Introduction

The semicircular canals and otolith organs detect angular and linear accelerations impinging on the head, respectively, and convey that information to the vestibular nuclei in the brainstem. The vestibular nuclei, in turn, give rise to effector pathways that participate in the control of eye movements, posture, balance, and some aspects of autonomic function (Vidal et al. 2015). The latter projections inform central presympathetic autonomic nervous system neurons influencing blood pressure, respiration and gastrointestinal activity about changes in posture, and body position with regard to gravity. Collectively, these functional pathways are referred to as the vestibulo-sympathetic reflex (VSR) (Yates et al. 2014).

One component of the VSR is a direct projection from the caudal vestibular nuclei to the rostral ventrolateral medulla (RVLM) (Holstein et al. 2011a), a region that contains the C1 adrenergic cell group as well as non-C1 presympathetic neurons (Ruggiero et al. 1994; Lipski et al. 1995; Schreihofer and Guyenet 1997). The RVLM is an important site of action of anti-hypertensive drugs such as clonidine, moxonidine and rilmenidine (Ernsberger et al. 1993; Regunathan and Reis 1996; Eglen et al. 1998) that bind to imidazoline receptors (IRs) and/or α2 adrenergic receptors in the region. While both C1 and non-C1 bulbospinal neurons may have a role in the medullary circuits regulating blood pressure, the rapid excitatory pathway from RVLM to spinal cord sympathetic preganglionic neurons appears to be mediated by glutamate (Deuchars et al. 1995) (reviews: Morrison 2003; Stornetta 2009).

Areas of the vestibular nuclei also project to the caudal ventrolateral medulla (CVLM), a region that participates in baroreflex-mediated modulation of blood pressure. CVLM receives baroreceptor-related inputs from the solitary nucleus and supplies both excitatory and GABAergic inhibitory projections to RVLM. We have previously demonstrated that the VSR pathways to RVLM and CVLM are activated by sinusoidal galvanic vestibular stimulation (sGVS) (Holstein et al. 2012), causing alterations in blood pressure (Cohen et al. 2013). Many of the vestibular nuclear neurons that are activated by this vestibular stimulus and that project to RVLM and/or CVLM display intense glutamate immunofluorescence (Holstein et al. 2016).

Imidazole-4-acetic acid-ribotide (IAARP) is a putative neurotransmitter/modulator in the CNS and an endogenous IR agonist (Prell et al. 2004). It is an intrinsic regulator of sympathetic drive, notably systemic blood pressure. IAARP is present in some central vestibular neurons (Martinelli et al. 2007) and in scattered cells throughout the CNS (Friedrich et al. 2007), including the RVLM (Holstein et al. 2011b). In the hippocampus, IAARP suppresses excitatory glutamatergic synaptic transmission at Schaffer collateral synapses with CA1 pyramidal cells (Bozdagi et al. 2011), and this inhibition includes both pre-and postsynaptic IR-mediated effects.

The goal of the present study was to determine whether vestibular neurons that participate in the VSR contain IAARP. Multiple-label immunofluorescence staining was conducted to identify IAARP in vestibular neurons that accumulated cFos protein in response to sGVS activation in animals that had received retrograde tracer injections in RVLM or CVLM. The additional goal of this study was to assess the co-localization of IAARP with intense glutamate immunofluorescence in VSR pathway neurons. The results demonstrate that IAARP is present in vestibular nucleus neurons of the VSR pathway, where it partially but not exclusively co-localizes with intense glutamate immunofluorescence.

Methods

Subjects

All experiments were conducted in accordance with the NRC Guide for the Care and Use of Laboratory Animals (8th Edition 2011) and were approved by the Institutional Care and Use Committee of the Icahn School of Medicine at Mount Sinai. Data were obtained from 14 adult male Long-Evans rats (Harlan Laboratories, MA) weighing 350–450 g. Seven of these animals received injections of the retrograde tracer FluoroGold (Fluorochrome, LLC, Denver, CO) in RVLM, five had FluoroGold injections in CVLM, and two received vestibular stimulation but no tracer injection. Sections from many (but not all) of these rats have been used in several of our previous studies (Holstein et al. 2014, 2016).

Retrograde tracer injections

FluoroGold was utilized because of its high sensitivity, low probability of uptake by fibers of passage, and lack of concomitant anterograde transport (Raju and Smith 2006; Schofield 2008). As previously described (Holstein et al. 2014, 2016), rats were anesthetized with isoflurane (4 % induction, 2 % maintenance), shaved and then placed in a computer-assisted stereotaxic frame (Leica Angle Two, Leica Microsystems, St. Louis, MO). Body temperature was maintained by laying the animal on a homeothermic pad regulated by feedback from a rectal thermometer. Ophthalmic ointment kept the eyes moist, and a single dose of analgesic (Buprenex, 0.05 mg/kg, SQ; Reckitt Benckiser Pharmaceuticals; Richmond VA) was administered preemptively. After draping for aseptic surgery, the head and neck were disinfected with Povidone. A midline incision was made from the top of the skull to the C1 vertebra, and the atlanto-occipital membrane was exposed. A glass pipette (tip OD 20–25 µm) filled with 2 % FluoroGold dissolved in saline was mounted on the Angle Two dorsoventral (DV) drive tilted 45° above the horizontal plane. The pipette tip was positioned at Bregma, and the target coordinates for RVLM (ML ± 2.34; AP −12.24; DV −10.21) or CVLM (ML ± 2.2; AP −12.80; DV −10.0) (Paxinos and Watson 2009) were entered in the computer. The Angle Two AP and ML drives were then adjusted, and the pipette was advanced under computer guidance toward the brainstem target via a small (~2 mm dia.) burr hole drilled at ML ± 2.3 mm. The tracer solution was iontophoresed at +5 µA for 10 min (7 s on, 7 s off), and the pipette was left in place for 2–3 min after the iontophoresis, before being slowly withdrawn. The neck muscles were sutured, and the skin was closed with surgical clips. Rats received 3–4 ml of sterile saline SQ at the end of the procedure. Analgesics were administered twice daily for 3 days after surgery (Buprenex; 0.05 mg/kg; SQ), and the animals were allowed to recover for 10–14 days before the terminal experiment was performed. The rats recovered fully, and we did not observe postoperative head tilt, body tilt, or locomotor activity favoring the unoperated side.

sGVS

Animals were anesthetized with isoflurane (4 % induction; 2 % maintenance) and maintained on a homeothermic pad. Ag/AgCl needle electrodes (BAK) connected to a computer-controlled current stimulator (Holstein et al. 2012) were inserted bilaterally under the skin over the mastoid processes. An individual sGVS stimulus comprised 5 cycles of binaural current (2 mA, 0.025 Hz). This stimulus was repeated 5 times with 3 min between repetitions. Rats were then allowed to recover from the anesthesia and were euthanized 90 min after the cessation of the last sGVS stimulus.

Tissue harvesting and processing

Perfusion, fixation, and tissue sectioning

Rats were anesthetized with isoflurane (as above) and then perfused transcardially 90 min after the sGVS stimulation, a time point of significant cFos protein accumulation in vestibular neurons after sGVS or tilt (Holstein et al. 2012, 2014). The initial perfusion with 100 ml of 37 °C 10 mM phosphate buffered saline (PBS) was followed by 500 ml of 4 % paraformaldehyde/0.2 % glutaraldehyde fixative in 0.1 M PB (pH 7.4) at room temperature. Brains were harvested immediately after perfusion, cut into blocks using an adult rat brain coronal matrix (Ted Pella, Inc.; Redding, CA), and stored at 4 °C in PBS with 0.02 % NaN3. The block containing the vestibular nuclear complex and ventrolateral medullary region was cut by vibrating microtome into 50 µm serial sections (~120 per animal) that were stored at 4 °C in PBS containing 0.02 % NaN3.

Anatomical localization

The locations of the four principal vestibular nuclei were determined in each tissue section by comparing the anatomical landmarks on the dorsal aspect of the tissue with a standard stereotaxic atlas (Paxinos and Watson 2005). Although this atlas was made using sections from Wistar rats, and the present study utilized the Long-Evans strain, we did not observe any significant neuroanatomical differences in the caudal vestibular nuclei, RVLM, or CVLM of the two strains. Moreover, the Angle Two injection guidance system utilizes Paxinos and Watson coordinates, and our success rate in targeting the structures of interest is ~80 %. This suggests that the Paxinos and Watson stereotaxic atlas coordinates are sufficient for the level of resolution required in this study.

The boundaries of RVLM and CVLM were determined by comparing the structures on the ventral aspect of the tissue sections with published maps and atlases of those regions (Paxinos and Watson 2005, 2009; Card et al. 2006; Bourassa et al. 2009; Goodchild and Moon 2009). Based on the most conservative estimates from these maps, RVLM was identified as a 1 mm rostrocaudal region extending from approximately 11.8–12.8 mm caudal to Bregma. The other dimensions of RVLM were determined by a triangle with nucleus ambiguus (pars compacta) at the apex and the ventral aspect of the medulla 1.4 and 2.2 mm lateral to the midline as the other two points. The CVLM region, located 12.8–13.6 mm caudal to Bregma, was similarly defined according to anatomical coordinates. Both regions correspond well with functional and physiological maps of RVLM and CVLM (Goodchild and Moon 2009). The locations of RVLM and CVLM were further verified in our experimental animals using tyrosine hydroxylase and GABA immunostaining of representative sections through these regions (Holstein et al. 2011a).

Primary antibodies

Individual tissue sections were immunostained using one of the following primary antibody combinations: (1) anti-FluoroGold and anti-IAARP; (2) anti-FluoroGold, anti-IAARP, and anti-cFos; (3) anti-FluoroGold, anti-IAARP, and anti-glutamate; and (4) anti-FluoroGold, anti-IAARP, anti-glutamate, and anti-cFos. Antibody sources and descriptions are shown in Table 1. The rationale for their use, staining protocols, and controls are described below.

Table 1 Primary antibodies used in this study
Full size table

In previous studies, we assessed several commercial antibodies against cFos including two unconjugated rabbit polyclonal antisera (Santa Cruz Biotechnology, Cat.# sc-253; Calbiochem, Cat.# PC38) and one rabbit polyclonal antiserum directly conjugated with AlexaFluor 488 (Santa Cruz Biotechnology, Cat. # sc-253 AF488) (Holstein et al. 2012). Although the three reagents stained the same vestibular regions and cell types, the unlabeled polyclonal sera provided more robust labeling in multiple-label immunofluorescence experiments. To assess non-specific staining, control tissue sections were exposed to a mixture of the rabbit polyclonal antibody preabsorbed with blocking peptide (Santa Cruz Biotechnology; Cat.# sc-253P); no immunolabeling was apparent in these sections. Previous studies in our laboratory have demonstrated that cFos-peroxidase immunolabeling in mock (non)-sGVS-stimulated rats comprises less than three neurons/vestibular region/tissue section (Holstein et al. 2012). Given the higher sensitivity of the peroxidase labeling system, fewer non-stimulation-dependent cFos-positive vestibular neurons are likely to be obtained using the immunofluorescence approach taken in the present study.

The FluoroGold tracer was amplified by immunostaining with a rabbit polyclonal anti-FluoroGold serum (Millipore; Cat.# AB153). Both the cFos and FluoroGold primary antibodies provided staining of the same vestibular regions and cells observed previously (Holstein et al. 2012, 2014, 2016).

The anti-IAARP sera were raised in rabbits immunized with an I-4-AA-RP-keyhole limpet hemocyanin conjugate (Prell et al. 2004). The production, characterization, and binding specificity of this antibody have been described in detail previously (Prell et al. 2004; Friedrich et al. 2007; Martinelli et al. 2007). The mouse IgG1 monoclonal anti-glutamate antibody was also produced in our laboratory. The complete description of the production, characterization, and specificity of this antibody was previously published (Holstein et al. 2004). Subsequently, we performed additional assays to verify (i) that glutamate can be fixed in situ by glutaraldehyde and (ii) that paraformaldehyde/glutaraldehyde-fixed glutamate can specifically be recognized by the monoclonal antibody (MAb 215B2) used in the present study (Holstein et al. 2011a).

Peroxidase/diaminobenzidine immunocytochemistry

To identify the FluoroGold injection sites in the RVLM or CVLM, regularly spaced vibratome sections (250 µm apart) from each injected rat were incubated in blocking buffer (PBS with 10 % normal goat serum, 0.02 % NaN3, and 0.1 % Triton X-100) for 4–6 h or overnight, and then in anti-FluoroGold antibody (1:5000 in blocking buffer) for 12–18 h. After thorough rinsing (6 changes of PBS over 4–6 h), the free-floating sections were incubated in peroxidase-conjugated goat anti-rabbit secondary antibody (Jackson ImmunoResearch Cat. # 111-035-144; 1:2000 in blocking buffer without NaN3) overnight. Sections were then rinsed in PBS (6 changes over 2 h) and incubated in diaminobenzidine (DAB; 0.5 mg/ml; Sigma D-5905; St. Louis, MO) diluted in 0.1 M Tris buffer (pH 7.6) with 0.01 % H2O2 for 5–10 min at room temperature. The reaction was stopped by repeated PBS rinses. Control sections were processed concomitantly and were treated as above except for the omission of primary and/or secondary antibodies (two sections per rat per control condition; 6 DAB control sections per rat).

Immunofluorescence

Tissue sections were processed for multiple-label immunofluorescence detection of combinations of primary and secondary reagents. Some sections were also stained with DAPI (ThermoFisher Scientific; Cat # D-1306) to visualize the location and approximate size range of cell nuclei (data not shown). Since the cFos, FluoroGold, and IAARP primary antibodies were raised in rabbit, the studies were performed by sequential application of primary and secondary antibodies. Briefly: (i) sections were exposed initially to rabbit anti-cFos primary followed by fluorochrome-labeled (e.g., DyLight 649) Fab fragment of a goat anti-rabbit IgG secondary, (ii) unreacted sites on the anti-cFos were blocked using a high concentration of unlabeled Fab fragment of goat anti-rabbit secondary antibody, (iii) stained sections were exposed to rabbit anti-FluoroGold followed by goat anti-rabbit IgG secondary antibody tagged with a different fluorochrome (e.g., DyLight 594), (iv) unreacted sites were blocked as above, and (v) the stained sections were exposed to rabbit anti-IAARP followed by goat anti-rabbit IgG secondary antibody tagged with a third fluorochrome (e.g., AlexaFluor 488). The sequence of primaries was varied across staining experiments, and each sequence was utilized in multiple tissue sections from each animal. An example of the overall staining procedure is provided below.

Example staining procedure: All steps were performed at room temperature with agitation on an orbital shaker. Free-floating sections were immersed in blocking buffer (3–6 h); incubated in rabbit anti-cFos primary antibody (1:500 in blocking buffer; 12–18 h); rinsed with multiple changes of PBS (4–8 h); incubated in AlexaFluor- or DyLight-conjugated goat anti-rabbit IgG (H + L) antibody Fab fragment (8 µg/ml in blocking buffer; Jackson ImmunoResearch; 12–18 h); rinsed in multiple changes of PBS (4–8 h); fixed with 2 % paraformaldehyde (10 min); rinsed (as above); incubated in unlabeled Fab fragment goat anti-rabbit IgG (20 µg/ml in blocking buffer; Jackson ImmunoResearch; 12–18 h); rinsed; incubated in a mixture of rabbit anti-FluoroGold primary antibody (1:400 in blocking buffer) and mouse anti-glutamate (1:10 hybridoma supernatant in blocking buffer) for 12–18 h; rinsed; and immersed in AlexaFluor-conjugated goat anti-rabbit IgG (H + L) (1:400 in blocking buffer; Invitrogen), and AlexaFluor-conjugated goat anti-mouse IgG (H + L) (1:300 in blocking buffer; Invitrogen). Following this staining, sections were incubated in DAPI solution (300 nM in PBS; 30 min). After final washes, all sections were mounted on glass slides and coverslipped using Prolong Gold antifade mounting medium (Invitrogen). Because different secondary antibodies have different degrees of sensitivity, various secondary antibodies and several alternative fluorochromes (in different sections) were used to detect each primary antibody.

Multiple staining controls were included in each experiment. To control for non-specific secondary antibody staining, sections were labeled with multiple secondary antibodies but no primary antibody. To control for secondary antibody cross-reactivity, tissue sections from each rat were labeled with multiple secondary antibodies but only one primary antibody. Lastly, to control for tissue autofluorescence, sections were treated with the blocking and rinsing steps, but no primary or secondary antibody incubations. All data for this study derived from experiments in which each secondary antibody bound exclusively to its appropriate primary antibody, and there was no binding to the inappropriate primary reagents. Importantly, all of the immunofluorescence-stained sections contained profiles that were stained by only one of the colors in the specimen’s secondaries mixture (see Figures). This provided further assurance within each individual section that only one secondary antibody bound to each primary antibody, and therefore that secondary antibody cross-reactivity was insignificant.

Microscopy and figure preparation

Sections were examined and imaged using a Zeiss Axioplan2 microscope equipped with structured illumination (ApoTome). Each session included images of intensity-calibrated fluorescent beads (InSpeck, ThermoFisher), which allowed the intensity values in the images to be transformed to a scale based on the beads and therefore reasonably independent of image recording conditions. This strategy is advantageous because it permits comparisons of fluorescence intensities across specimens, experiments, and imaging sessions. Using this approach, the level of fluorescence designated as “intense” was set at 75 % of maximal fluorescence intensity.

Immunofluorescence was recorded using multichannel gray scale single slices and z-stacks. The images were analyzed by assessing the immunofluorescence present in each gray scale optical slice of each single channel separately. For z-stack images, antibody co-localization was initially screened using maximum intensity projections through the z-stacks, and both x–z and y–z side views of pseudocolored image stacks constructed using the Zeiss ApoTome software. Co-localization was then verified by comparing the gray scale immunofluorescence present in each individual channel within the same optical slice. Volocity, ImageJ and AxioVision were used to generate z-axis maximum intensity projections of whole stacks or selected subsets of optical sections for presentation figures. Figures were prepared (cropped, sized, colored, labeled, annotated) using Adobe Photoshop and Illustrator CC2015. Adjustments of brightness and contrast were performed using the Photoshop levels and curves tools applied to all parts of each image. The details of our image processing strategy for figure preparation, albeit with earlier versions of the same software, are published in (Holstein et al. 2014).

Results

Unilateral FluoroGold injections into RVLM (N = 7) and CVLM (N = 5) were performed 10–14 days prior to euthanasia. This provided sufficient time for the tracer to be transported retrogradely to the cell bodies of vestibular nuclear neurons. After tissue sectioning, the injection sites and diffusion penumbrae were visualized in regularly spaced sections (one every 250 µm) through the ventrolateral medulla stained for FluoroGold immunocytochemistry. The rostrocaudal extent of the tracer in each rat used in the present study is presented in Table 2. In general, the injection sites and surrounding penumbrae extended less than 1 mm medio-laterally and less than 750 µm dorso-ventrally.

Table 2 FluoroGold injection sites
Full size table

Additional sets of transverse sections through the vestibular nuclei from animals with verified tracer injections were then double labeled with anti-FluoroGold and anti-IAARP. Double-labeled neurons were observed in the caudal half of the medial and spinal vestibular nuclei (MVN and SpVN, respectively), but not more rostrally and not in magnocellular MVN. Although double-labeled neurons were present bilaterally, they were more prevalent in vestibular nuclei ipsilateral to the injection site. Three cytological types of FluoroGold-filled and IAARP-immunofluorescent neurons were observed: small fusiform neurons, larger multipolar cells, and small spherical neurons (Fig. 1).

Fig. 1
figure 1

Examples of the three types of FluoroGold-labeled neurons (punctate label) that co-localize IAARP. a A fusiform neuron in SpVN that was retrogradely filled with tracer after an injection in ipsilateral RVLM. b A larger diameter neuron in caudal MVN that was double labeled after an injection in ipsilateral CVLM. c A small spherical neuron in the MVN that was double labeled after a tracer injection into the contralateral RVLM. Scale bar in C is 10 µm in all panels

Full size image

The presence of FluoroGold in vestibular cell bodies was interpreted as evidence that the tracer-filled neurons sent axonal projections to the ventrolateral medulla. To verify that such cells participate in the functional VSR, low frequency binaural sGVS was used to stimulate vestibular afferent nerve fibers in the rats prior to euthanasia. Activation of central vestibular neurons was then assessed by immunofluorescence detection of cFos protein (Holstein et al. 2012). Tissue sections from the tracer-injected rats were immunostained concomitantly for cFos, FluoroGold, and IAARP. We found that IAARP was present in sGVS-activated vestibulo-sympathetic projection neurons, as well as in vestibular neurons that were not tracer-filled and/or did not accumulate cFos protein after sGVS (Fig. 2). The activated (cFos+) VN neurons that contain IAARP but not FluoroGold (e.g., Fig 3b) are interpreted as vestibular neurons with projections exclusively to the contralateral R/CVLM, or VSR cells with projections to the ipsilateral side, but to the non-injected VLM subregion (CVLM or RVLM). In both circumstances, cFos activation is expected, but retrograde tracer is not. Triple-labeled cells of all three cytologic types were observed in both SpVN and MVN, projecting to RVLM and CVLM (Fig. 3).

Fig. 2
figure 2

Overview of vestibular neurons with direct projections to the ventrolateral medulla (FluoroGold-immunopositive), which were activated by sGVS (cFos-positive) and co-localized IAARP. Top panel shows a cluster of single-, double-, and triple-labeled neurons in ipsilateral SpVN from a rat with a FluoroGold injection in RVLM. The individual channel images of the area delineated by the dashed rectangle are shown in the panels below. These single channel images illustrate co-localization of cFos and FG in two neurons, the activation of one additional neuron that was not retrogradely filled, and the presence of IAARP in all three of those cells, as well as others in the field. Scale bar is 10 µm

Full size image
Fig. 3
figure 3

IAARP is present in the cell bodies of sGVS-activated vestibular neurons with direct projections to the ventrolateral medulla. Each large panel (a–c) illustrates one triple-immunofluorescent vestibular neuron, as well as several double-and single-labeled cells. The side panels are single-channel images of the triple-labeled neurons in the larger panels to the left. For each set of side panels, FluoroGold label is shown at the top, cFos in the center, and IAARP at the bottom. a A multipolar neuron in SpVN with projections to contralateral CVLM. b A fusiform neuron in MVN that projects to ipsilateral RVLM. The smaller neuron in the upper left quadrant of the panel is cFos and IAARP-labeled, but did not contain retrograde tracer. c A small spherical neuron in SpVN with projections to ipsilateral RVLM. Scale bar is in C is 10 µm for all large panels. The single channel images are slightly enlarged

Full size image

We have previously reported that many sGVS-activated vestibular neurons with projections to CVLM or RVLM are intensely glutamate immunofluorescent (Holstein et al. 2016). To determine whether IAARP is co-localized in these neurons, sections from FluoroGold-injected rats were immunostained to visualize the tracer together with IAARP and glutamate. The low-affinity monoclonal anti-glutamate antibody used in this study has previously been described in detail (Holstein et al. 2004, 2011a). While some neurons had barely detectable immunofluorescence signal, presumably reflecting the metabolic role of glutamate in cells, a subset of neurons were intensely immunofluorescent. Only these latter cells were designated as glutamate-immunofluorescent neurons in the present study (see “Methods” section). We observed that many, but not all, of the neurons that co-localized tracer and IAARP were glutamate immunofluorescent (Fig. 4). All three cytological types of neurons that were previously identified by retrograde tracing alone (multipolar neurons, fusiform neurons, and small globular neurons) displayed triple-label co-localization. These cells were present in both SpVN and in MVN, and projected either ipsilaterally or contralaterally. It is noteworthy that some activated VSR neurons (cFos- and FluoroGold-labeled) did not appear to be glutamate immunofluorescent. Although negative immunolabeling data must be interpreted conservatively, these glutamate-non-immunofluorescent neurons were observed in fields in which other neural elements displayed intense glutamate immunofluorescence, suggesting that antibody penetration and imaging conditions were not significant causes for the negative results. An example is shown in Fig. 5.

Fig. 4
figure 4

Examples from two rats of the distributions of IAARP-, glutamate- and FluoroGold-immunofluorescent neurons in the caudal vestibular nuclei. a Low magnification image of SpVN following a tracer injection into the contralateral RVLM. Neurons surrounded by single boxes displayed FluoroGold tracer, but not IAARP or glutamate immunofluorescence, when the individual channels were evaluated in gray scale. Neurons surrounded by two rectangles were FluoroGold and IAARP-immunofluorescent, but did not show intense glutamate label. Cells surrounded by three boxes were triple-labeled. b A higher magnification image of SpVN following a tracer injection in the ipsilateral CVLM. Neurons identified by two arrows co-localized IAARP and FluoroGold, but not intense glutamate immunofluorescence; neurons indicated by three arrows co-localized all three markers. Scale bars 50 µm in a; 10 µm in b

Full size image
Fig. 5
figure 5

Examples of IAARP- and FluoroGold-labeled vestibular neurons that are not intensely glutamate immunofluorescent. a The four FluoroGold-filled neurons in this image of SpVN (tracer injection in ipsilateral RVLM) are indicated by arrows. b The single channel image of IAARP in the same field. The arrows point to the same cells indicated in a. c The single channel image of glutamate immunofluorescence in the same field. The arrows point to the same cells indicated in a and b. None of the four neurons shows intense glutamate immunofluorescence, although a nearby cell body (asterisk) and numerous processes are labeled. Scale bar in a 20 µm

Full size image

Of 223 FG+ and IAARP+ neurons examined in this study, 150 (67 %) were located in SpVN and 73 (33 %) were observed in caudal MVN. With the simplifying assumption that all VSR neurons project only ipsilaterally or contralaterally and that there are no bilaterally projecting VSR neurons, 151 of the 223 labeled cells (68 %) projected ipsilaterally, and 72 (32 %) projected contralaterally. Of these, 102 (46 %) IAARP+ cells were located in SpVN and projected ipsilaterally, 48 (22 %) IAARP+ cells were located in SpVN and projected contralaterally, 49 (22 %) IAARP+ cells were located in caudal MVN and projected ipsilaterally, and 24 (11 %) IAARP+ cells were located in caudal MVN and projected ipsilaterally. Of the smaller subset of these cells stained for FG, IAARP, and glutamate (N = 89), 52 (58 %) showed intense glutamate immunofluorescence and 37 (42 %) showed no appreciable glutamate immunolabel in microscope fields where adjacent non-VSR neurons were glutamate immunofluorescent.

To confirm the presence of IAARP in the functional VSR pathway, some sections from tracer-injected and sGVS-stimulated animals were immunolabeled for FluoroGold, cFos, IAARP, and glutamate. Quadruple-labeled neurons of all three cytological types were observed in these sections (Fig. 6), providing strong evidence that IAARP are present in most cells of the glutamatergic limb of the VSR.

Fig. 6
figure 6

Co-localization of IAARP and glutamate in sGVS-activated (cFos-positive) and retrogradely filled (FluoroGold-positive) vestibular neurons. The larger panels (a–c) illustrate quadruple-labeled neurons of three cytologic types. The side panels are the single channel images of the labeled neurons in the larger panels to the left. For each set of side panels, FluoroGold tracer is shown in the upper left panel, IAARP immunofluorescence is shown in the upper right panel, glutamate immunofluorescence is shown in the lower left panel, and cFos immunofluorescence is shown in the lower right panel. a A fusiform neuron in MVN after a tracer injection in the ipsilateral CVLM. b A large multipolar neuron in MVN after a tracer injection in the ipsilateral RVLM. c A small spherical neuron in SpVN after a tracer injection in ipsilateral CVLM. Scale bars in all three panels: 10 µm

Full size image

Discussion

The results of this study demonstrate that IAARP is present in vestibular nucleus neurons of the VSR pathway and frequently co-localizes in cells with intense glutamate immunofluorescence. The significance of these findings stems from the putative role of IAARP as an endogenous modulator of CNS neurons involved in blood pressure control.

CNS cells and pathways involved in blood pressure control are clinically important as drug targets for the treatment of patients with hypertension. Clonidine and several related drugs synthesized for that purpose contain an imidazole(in)e ring and are ligands at both a2 adrenergic receptors (a2ARs) and IRs (Bousquet et al. 1984; De Vos et al. 1994; Tolentino-Silva et al. 2000). These compounds are thought to exert their hypotensive action through inhibition of tonically active presympathetic neurons in RVLM (Karppanen 1977; Bousquet et al. 1984; Atlas 1991). The binding sites for these drugs have not been identified definitively, but three IR subtypes (I1-, I2- and I3-Rs) have been distinguished pharmacologically (Eglen et al. 1998; Head and Mayorov 2006). I1Rs are found in highest density in the brainstem, primarily in RVLM and CVLM (Moreira et al. 2004), are localized to plasma membranes, are principally neuronal, and have high affinity for the anti-hypertensive drugs containing a synthetic imidazoline ring (Bousquet et al. 1984; Ernsberger et al. 1990; De Vos et al. 1994; Ernsberger and Haxhiu 1997; Wang et al. 2003; Chan et al. 2005; Nikolic and Agbaba 2012). In addition, several studies have suggested that I1Rs are coupled to G-protein-mediated signal transduction pathways, resulting in the activation of phosphatidylcholine-sensitive phospholipase C (Separovic et al. 1997), increased phosphorylation of mitogen-activated protein kinases (MAPK1 and MAPK3) (Zhang and Abdel-Rahman 2005), and inhibition of adenylate cyclase (Greney et al. 2000). Activation of any of these second messenger mechanisms in RVLM bulbospinal neurons would result in a decrease in systemic blood pressure (for review, see Del Bello et al. 2015). I2Rs appear to be mainly allosteric binding sites on monoamine oxidases and possibly other proteins (Remaury et al. 2000), are distributed throughout the brain, have been associated with glial cells as well as neurons (Lione et al. 1998), and are thought to be involved primarily in psychiatric disorders, analgesia, opiate withdrawal, and Parkinson’s and Alzheimer’s diseases (Garcia-Sevilla et al. 1990; Parini et al. 1996; Eglen et al. 1998). I3Rs mediate insulin secretion in the pancreas (Morgan 1999) and an I3-like binding site may also participate in blood pressure modulation (Prell et al. 2004), possibly underlying the association between hypertension and diabetes (Bousquet et al. 1984; Chan 1998; Morgan 1999).

Based on the clinical efficacy of I1R synthetic agonists, investigators have sought to identify the endogenous ligands that bind with high affinity to these sites in brain. Four putative endogenous I1R ligands have been proposed to date: harmane, agmatine, clonidine-displacing substance, and IAARP (reviews: Reis and Regunathan 2000; Halaris and Piletz 2007; Wu et al. 2008; Ghazaleh et al. 2015). IAARP displays high affinity binding to I1Rs and possible also I3Rs, producing physiological effects that are blocked by the appropriate antagonists (Prell et al. 2004). Biochemical assays and immunolabeling studies have demonstrated that IAARP is present in neurons and processes in the RVLM, and microinjection of IAARP into the RVLM region alters blood pressure in rodents (Prell et al. 2004). Results of the present study demonstrate that IAARP is localized in vestibular neurons that send direct projections to RVLM or CVLM and that are activated by vestibular stimulation. Thus, the results demonstrate that one important source of IAARP in RVLM is the VSR pathway that modulates blood pressure in response to changes in posture.

The second key finding in the present study is that IAARP- and glutamate immunofluorescence frequently co-localize within individual VSR neurons. Electrophysiological studies in rat hippocampus have examined the interaction between glutamatergic and IAARP-mediated synaptic effects. The hippocampus contains high densities of both IAARP-immunopositive neurons (Friedrich et al. 2007) and IR binding sites (Piletz et al. 2000). Using extracellular and whole-cell recordings of the Schaffer collateral-commissural pathway in rat hippocampal slices, IAARP was found to bind to IRs, resulting in the inhibition of excitatory transmission from Schaffer collaterals to CA1 pyramidal neurons. Moreover, the frequency of miniature excitatory postsynaptic currents was decreased and paired-pulse facilitation increased after bath application of IAARP, suggesting that the site of action was primarily presynaptic (Bozdagi et al. 2011). These studies demonstrate that IAARP binding to I1Rs occurs both pre- and postsynaptically, resulting in suppression of glutamatergic neurotransmission. It is noteworthy in this context that autonomic dysregulation, including irrecoverable hypotension due to abnormalities in the hypoxia-sensitive regions of the CA1 and CA4 hippocampal fields, is speculated to play an important role in sudden unexpected death in epilepsy (Bozorgi et al. 2013) and in sudden infant death syndrome (Harper and Kinney 2010; Machaalani and Waters 2014).

Conceivably, the co-localization of IAARP- and intense glutamate immunofluorescence in VSR neurons represents an efficient configuration for up- and down-modulation of RVLM presympathetic neuronal activity. Co-release of the two neuroactive agents from single VSR pathway axons would activate fast ionotropic glutamate receptors located postsynaptically on RVLM neurons, and both pre- and postsynaptic IRs. Since the I1R is the most likely binding target and is a G-protein-coupled receptor, this slower transmission would allow a brief burst of excitatory activity to activate RVLM bulbospinal neurons before the IAARP release would cause both presynaptic inhibition of further co-transmitter release from the axon terminal and postsynaptic inhibition of the recipient RVLM neuron.

In fact, studies assessing the effects of synthetic imidazoline agonists acting in the reticular formation on sympathetic outflow and cardiovascular function have also suggested that IRs are located presynaptically (Head 1999; Chan et al. 2005, 2007). It has been proposed that I1Rs and a2ARs are organized in series in the RVLM such that IRs are proximal to presynaptic sympatho-inhibitory noradrenergic terminals (Molderings and Göthert 1999), and α2ARs lie more distally, possibly as somatic autoreceptors (Head 1999; Chan et al. 2005). The present results suggest further complexity to the receptor interplay in RVLM.

This complexity is underscored by the observation that some VSR pathway neurons appear to utilize GABA (Holstein et al. 2016), rather than an excitatory glutamatergic neurotransmitter. It is possible that some of the non-glutamate immunofluorescent and IAARP-positive VSR neurons observed in the present study are part of the GABAergic limb of the VSR, which is primarily directed to CVLM. In this context, it is interesting to note that an electrophysiological study in acute slices of rat striatum provided evidence that presynaptic I1Rs are involved in the inhibition of GABAA-mediated inhibitory postsynaptic currents by the imidazoline drug moxonidine (Tanabe et al. 2006). In fact, we have presented preliminary data showing that IAARP has similar inhibitory effects on inhibitory postsynaptic currents in striatal slices (Artis et al. 2007). Postsynaptic effects are also likely to contribute to shaping the inhibitory synaptic response to IAARP obtained in the acute hippocampal slices, and additional studies will be necessary to identify the mechanisms underlying these effects in the RVLM.