Involvement of median preoptic nucleus and medullary noradrenergic neurons in cardiovascular and sympathetic responses of hemorrhagic rats

The infusion of hypertonic saline solution (HSS) is known to be beneficial to the treatment of hypovolemic hemorrhage (HH). The central mechanism of HSS-induced cardiovascular and autonomic recovery of animals subjected to HH remains unclear. Hence, the present study evaluated the involvement of median preoptic nucleus (MnPO) and medullary noradrenergic neurons (A1 and A2) in HSS-induced cardiovascular and sympathetic responses in hemorrhagic rats. The wistar rats were subjected to specific lesion of noradrenergic neurons through the nanoinjections of anti-DβH-saporin into caudal ventrolateral medulla (A1 neurons) and nucleus of the solitary tract (A2 neurons). After recovery, mean arterial pressure (MAP) and renal sympathetic nervous activity were recorded. The HH was performed through blood withdrawal until a MAP of 60 mmHg was attained. In sham rats, HSS infusion (3M NaCl) reestablished MAP without change in HH-induced sympathoinhibition. The muscimol (agonist of GABAA receptor) was nanoinjected in MnPO during HH and MnPO inhibition abolished the recovery of MAP and HSS-induced sympathoinhibition. Simultaneous lesions of A1 and A2 abolished MAP restoration and sympathoinhibition after HSS infusion. These results suggest that the recovery of MAP and HSS-induced sympathoinhibition in hemorrhaged rats depend on intact neural projections from A1 and A2 to MnPO.


Nanoinjection into MnPO and pharmacological inhibition. The muscimol or saline nanoinjected
into MnPO at an interval of 10 minutes after HH. The nanoinjections were performed with a glass micropipette coupled to a manual pressure system. The parietal and frontal bone were partly removed to perform the nanoinjections. The bregma was located and the coordinates were recorded for positioning of the glass pipette. The injected volume of each solution was controlled by meniscus displacement of the solution in the micropipette through a surgical microscope equipped with an ocular lens with calibrated reticulum.
In order to achieve pharmacological inhibition, the animals received nanoinjections of Muscimol (GABAA receptor agonist, 4 mM, 100 nL, Sigma-Aldrich, St. Louis, MO, USA) in the MnPO. The sham group received nanoinjections of saline (NaCl, 0.15 M, 100 nL) were performed in the MnPO. At the end of the experiments, the Evans blue dye (4%, 100 nL, Sigma-Aldrich, St. Louis, MO, USA) was nanoinjected into the same region for histological confirmation of the sites. The Paxinos & Watson atlas 16 was used to determine the coordinates. For all nanoinjections into the MnPO, a glass micropipette was positioned at 0.6 mm rostral to bregma, 0.0 mm lateral to midline, and 7.2 mm below dura mater 17 . Only animals whose nanoinjections were restricted to the MnPO region were considered for analysis.

Nanoinjections into the CVLM and NTS.
The animals belonging to the A1 and A2 neuronal lesion protocol were anesthetized with a mixture of ketamine (10%, 1 mL·kg −1 , i.p., Syntec, Santana de Parnaíba, SP, Brazil) and Xylazine (2%, 0.7 mL·kg −1 , i.p., Syntec Santana de Parnaíba, SP, Brazil) on a stereotaxic apparatus (Insight Ltda., Ribeirão Preto, SP, Brazil) with incisor bar 11 mm below the interaural line. An incision was made in the posterior inferior region of the head, to expose the occipital bone and the atlanto-occipital membrane. The occipital bone was removed and the calamus scriptorius was used as a reference point for the stereotactic coordinates. In order to achieve A1 and/or A2 neuronal lesions, the anti-dopamine-β-hydroxylase-saporin complex (anti-DβH-saporin, 100 nL, 0.105 ng·nL −1 , Advanced Targeting Systems, San Diego, CA, USA) was nanoinjected into the CVLM and NTS region, respectively. In sham groups, the equimolar of Saporin (100 nL, 0.022 ng·nL −1 , Advanced Targeting Systems, San Diego, CA, USA) was nanoinjected into the same site. For all nanoinjections into the CVLM, a glass micropipette was positioned at 0.3 mm rostral and 0.2 mm caudal from the calamus scriptorius, 1.8 mm lateral from the mid-line, and 1.8 mm ventral from the dorsal surface. For all nanoinjections into the NTS, a glass micropipette was positioned at 0.0 and −0.5 mm caudal to calamus scriptorius, 0.0 mm lateral from the mid-line and 0.3 mm ventral from the dorsal surface. These coordinates were based on the region of the CVLM and NTS consisting of the A1 and A2 neurons groups, respectively 18,19 .
At the end of the central nanoinjections, the incision was sutured with surgical line prior to the administration of analgesic (Flunixin, 0.02 mL·kg −1 , i.m., CHEMITEC, Brazil, SP). The animals were later housed during 20 days with free access to water and food to ensure surgical recovery and lesions establishment.
Surgical procedures. The animals were subjected to anesthetic induction through the administration of halothane (2%, Tanohalo, Cristália, Itapira, SP, Brazil) in 100% O 2 prior to the catheterization of femoral artery and vein. After vein catheterization, anesthesia was maintained by administration of urethane (1.2 g·kg −1 , i.v., Sigma-Aldrich, St. Louis, MO, USA). Additional catheter was inserted into the right carotid artery to withdraw blood during HH. Tracheostomy was performed to reduce airway resistance. In order to record renal sympathetic nervous activity (RSNA), left renal nerve was isolated and positioned on silver bipolar electrodes. The body temperature was maintained between 36 °C and 37 °C with a thermostatically controlled heated table.
Recording of renal sympathetic nerve activity (RSNA). The RSNA was recorded through the left renal nerve with bipolar silver electrodes. The renal nerve was located, dissected and covered with mineral oil (Nujol, Schering-Plough, São Paulo, SP, Brazil) prior to the placement of electrodes for recording. The signals were obtained using a high-impedance probe connected to the amplifier (P511, Grass Instruments, Quincy, MA, USA). The signal was amplified 20.000 times, digitized and band-pass filtered (30-1000 Hz). The nerve signal was recorded continuously (2000 samples · s −1 , PowerLab 4/25, ML845, ADInstruments, Bella Vista, Austrália) rectified and integrated at 1 s intervals using LabChart software (v.7.3.7., ADInstruments, Bella Vista, Austrália). At the end of each experiment, ganglionic blocker hexamethonium (30 mg·kg −1 , b.wt., i.v., Sigma-Aldrich, St. Louis, MO, USA) was administered to determine the background noise. The level of RSNA was expressed as a percentage of baseline after subtraction of the noise. Functionality of baroreceptor reflexes in baseline period was also evaluated. For this, fast Fourier transform was performed on 60 s windows after the AP and SNA signals were configured and sampled at 1000 Hz. The frequency resolution of the spectra was 0.2 Hz/bin. The spectra in this paper show frequencies between 0 and 10 Hz. The same windows were used for coherence analysis by using coherence script for Spikes 2 software as described 22 . Recording of blood flow. The renal blood flow (RBF) and aortic blood flow (ABF) baseline were recorded through a miniature probe that was placed around the left renal artery and abdominal aorta. Miniature probes were connected to a T206 flowmeter (Transonic Systems, Inc., Ithaca, NY, USA) which determines the flow rate in absolute values (ml·min −1 ). The signals obtained were transferred to the acquisition and data analysis software (PowerLab 4/25, ML845, ADInstruments, Colorado Springs, CO, USA). Data was digitized at a sampling frequency of 1000 samples per second. The hind limb blood flow (HBF) was calculated by using the following equation: (ABF) − (2 x RBF). Each of the fourth brainstem section were washed in ImmunoBuffer (IB, Triton 0.3% in phosphate buffer saline/PBS, Sigma-Aldrich, St. Louis, MO, EUA) followed by a 30-min incubation in 2% normal horse serum (Vector Laboratories Inc., Burlingame, CA, USA) in IB. The sections were incubated overnight with mouse monoclonal antibody (1:2000 dilution, Immuno Star Inc., Hudson, WI, USA) with 2% normal horse serum in IB, followed by an overnight incubation with biotinylated horse anti-mouse IgG (1:500 dilution, Vector Laboratories Inc., Burlingame, CA, USA). After these incubations, the sections were processed with the avidin-biotin procedure using Elite Vectastain reagents (Vector Laboratories Inc., Burlingame, CA, USA). The diaminobenzidine (DAB) was used to produce a brown cytoplasmic TH reaction product. Statistical analysis. The statistical analyses were performed using the GraphPad Prism software (v 5.01).

Immunohistochemistry.
Cardiovascular and autonomic parameters were expressed as mean ± standard error of the mean (SEM). The MAP, HR and RSNA variations were analyzed using two-way analysis of variance (ANOVA two way) followed by Tukey's post hoc test. The value of p < 0.05 was considered a significant difference.
The cell count was expressed as mean ± SEM. The number of cells counted for each section was compared by one-way analysis of variance (ANOVA one way) followed by Tukey's post hoc test. The total cell counts for groups A1, C1, A2, and C2 were compared between groups by Student's t-test. Value of p < 0.05 was considered to be significantly different.
Data availability statements. The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Effects of A1 and A2 simultaneous lesion on HSS-induced cardiovascular and sympathetic
responses. The quantification of TH positive neurons demonstrated that A1 + A2 sham rats (A1 + A2S; n = 6), which received saporin nanoinjections into CVLM and NTS regions, have an average of 28 cells per section at the CVLM region that encompassing the A1 neuronal cluster and an average 30 cells per section at the NTS region that encompasses A2 neuronal cluster. The average number of cells per section was reduced to 11 in the CVLM and 12 in the NTS in A1 + A2-lesioned rats (A1 + A2L; n = 6) which received anti-DβH-saporin nanoinjections. Hence, approximately 59% of A1 neurons and 61% of A2 neurons in the A1 + A2L group was reduced as compared to the A1 + A2S group (Fig. 4B). The anti-DβH-saporin nanoinjections in the CVLM and NTS regions did not promote significant changes in the number of TH positive cells at NTS region that encompasses C2 neuronal cluster (Fig. 4B, 2% in relation to A1 + A2S group) and in the VLM region, which includes the C1 neuronal cluster (Fig. 4B, C1: 11% as compared to A1 + A2S group).

Effects of MnPO inhibition on baroreceptor reflexes. Baseline functionality of baroreceptor reflexes
was evaluated by analysis of coherence between arterial pressure (AP) and RSNA. Similar results were observed in MnPOS and MnPOI groups. In both groups, RSNA power spectra displayed a prominent peak corresponding to the heart rate, indicating a prominent role of baroreceptor reflexes in modulating spontaneous RSNA discharge (Fig. 6). Similarly, correlation between RSNA bursts and cardiac cycle leads to high AP-RSNA coherence values (Mean -MnPOS: 0.84 ± 0.03 vs. MnPOI: 0.83 ± 0.03, Fig. 6, Table 2).

Discussion
Despite the establishment of cardiovascular response to HSS infusion in hemorrhagic animals 23,24 , the involvement of CNS pathways still remains unclear. In the present study, the participation of MnPO, A1 and A2 noradrenergic neurons in HSS-induced recovery during HH was demonstrated through; (I) The attenuation of MAP restoration and renal sympathoexcitation after HSS infusion and MnPO inhibition. (II) Non-alteration in the patterns of HSS-induced cardiovascular and sympathetic responses in sham rats with A1 or A2 lesion; (III) Simultaneous A1 and A2 lesion which abolished the HSS-induced MAP recovery and renal sympathoinhibition in sham rats. Taken together, these results are consistent with the view that medullary noradrenergic projection to MnPO are crucial to the development of reflex responses to the HSS-induced cardiovascular recovery during HH.
We demonstrated that there were no baseline changes in RBF, ABF and HBF in all analyzed groups. These results allow us to infer that the cardiovascular and autonomic changes observed are due to HH and HSS infusion and not due to baseline hemodynamic changes caused by previous neuronal lesions. We also demonstrated that A1 and A2 neuronal cluster simultaneous lesion and MnPO inhibition do not alter the functionality of baroreceptor reflexes. More specifically, this animals present: synchronization of RSNA to the cardiac cycle, a prominent peak in the power spectra corresponding to the heart rate, and high AP-SNA coherence. In this way, modulation of spontaneous RSNA discharge in the baseline period is mainly regulated by baroreceptor reflexes. Thus, the functionality of baroreceptor reflexes in these animals indicate once again that the results observed in the present study are not due to baseline hemodynamic changes.  Table 2. Baseline values of renal blood flow (RBF), aortic blood flow (ABF), hindlimb blood flow (HBF), and coherence between arteria pressure and renal sympathetic nervous activity (Coherence AP-RSNA). MnPOS (MnPO sham group), MnPOI (MnPO Inhibition group), A2S (A2 sham group), A2L (A2 lesioned group), A1S (A1 sham group), A1L (A1 lesioned group), A1S (A1 + A2 sham group), A1 + A2L (A1 + A2 lesioned group).
Previous studies have shown that rapid blood volume withdrawal during HH could lead to reductions in venous return, HR, cardiac output, fall in blood pressure 9,25 as well as biphasic RSNA response 26 . The activation of arterial baroreceptors mediates renal sympathoexcitation prior to vagal cardiopulmonary afferents-induced renal sympathoinhibition. Our results showed that the removal of approximately 15-25% of total blood volume promotes hypotension, bradycardia and reduction in RSNA. Thus, it is suggested that hypovolemia-induced MAP reduction is partly associated to the reduction in HR and cardiac output. The renal sympathoinhibition suggests a Moreover, our data showed that the HSS infusion reestablished MAP and HR of sham rats. Several studies have reported HSS infusion-induced hemodynamic improvement after HH 9,17 . Highlighting the sympathetic component influence in the hyperosmolarity-responses, the sympathoinhibition generated in the renal territory during HH was maintained after sodium overload. Hence, these data are in agreement with studies which showed renal sympathoinhibition and vasodilation induced by high plasma sodium concentration [27][28][29][30] The renal vasodilation may be essential for the maintenance of blood flow to the kidneys.
Like in the previous study 17 , HSS infusion did not promote plasma volume expansion. Hence, volume replacement seems insufficient to explain cardiovascular hyperosmolarity-induced recovery. The chemoreceptors, baroreceptors, aortic and carotid afferents are involved in HSS-induced cardiovascular recovery 9 . In addition, experimental evidence suggests that acute increases on plasma sodium concentration could trigger central components 31 .
The role of MnPO as the control center for the electrolytic balance and integration is well established. The MnPO receives dense excitatory, mostly glutamatergic nature, projections from central osmoreceptors 32,33 . These projections includes those of organum vasculosum of the lamina terminalis (OVLT) and the subfornical organ (SFO) 32,34,35 . The electrical stimulation of SFO often lead to an increase in the frequency of action potentials of MnPO neurons 36 . In addition, this nucleus controls the release of sympathetic activity, vasopressin, thirst and salt appetite 32,37,38 . In summary, MnPO can be considered as a regulatory center of body fluids homeostasis. The pharmacological inhibition of MnPO which abolished sodium overload-induced MAP restoration after HH is consistent with the previous findings. Furthermore, we showed that MnPO is not directly involved in the HR related responses, since the inhibition of this nucleus did not alter sodium overload-induced HR restoration.
We demonstrated sympathoexcitation in the renal territory after HSS infusion and MnPO inhibition. These data are in agreement with the previous MnPO inhibition that abolished high plasma osmolarity-induced renal vasodilation 17,39 . These new data suggest involvement of MnPO in cardiovascular and autonomic responses to HSS infusion after hypovolemia.
Having established the importance of MnPO to the sodium overload-induced cardiovascular and autonomic adjustments after HH, we investigated the involvement of a specific medullary region on these responses. Considering the neuroanatomical studies that have demonstrated dense projections of the A1 and A2 neuronal clusters to MnPO 11-13 , we hypothesized A1 and A2 neurons as potentials central components that regulate the reflex adjustments induced by HSS infusion in hemorrhagic rats.
The volume and composition of the circulating liquid are transmitted to the CNS through afferent fibers of peripheral sensors such as chemoreceptors, baroreceptors, cardiopulmonary receptors, and peripheral osmoreceptors 40 . In the NTS, these afferent fibers perform the first synapse and trigger a series of connections in the CNS 41 . From the NTS, A2 neuronal cluster are projected directly and indirectly, via A1 neuronal cluster, to the MnPO [11][12][13] . The dense projections of MnPO are sent to PVN. The PVN is a hypothalamic nucleus that regulates humoral and autonomic responses to hyperosmolarity, through vasopressin secretion and sympathetic preganglionic neurons projections, respectively 37,38 . Studies have shown that the PVN regulates the autonomic responses through its projections to the intermediate-lateral column (IML) and to the rostral ventrolateral medulla (RVLM), where sympathetic preganglionic neurons and sympathetic pre-motor neurons are located, respectively 37,40 . Thus, the integration of this extensive neural network culminates in autonomic, hemodynamic and endocrine reflexes in order to maintain body homeostasis in response to osmotic stimuli 28,30 . Within this neuronal pathway, the A1 and A2 noradrenergic neurons cluster highlight. Our studies characterize these structures as part of the fundamental pathways in the transmission of information about sodium overload-induced changes in plasma osmolarity after HH (Fig. 8). Although, the lesion of A1 or A2 neuronal clusters did not alter sodium overload -induced cardiovascular and autonomic responses after hypovolemia, we cannot exclude central parallel pathways which regulate same function. The latter assumption is supported by the previous studies which demonstrated that concomitant lesion, not the individual lesion, of AV3V region and NTS reduced MAP in spontaneously hypertensive rats 42 .This investigation demonstrated the importance of parallel pathways to cardiovascular control. In this sense, concomitant lesion of A1 and A2 neuronal cluster could cause complete loss of compensatory mechanisms. It is important to note that A1 and A2 neuronal clusters have the same embryological origin, similar projections to the hypothalamic regions and similar responses to hyperosmotic stimuli. Hence, it is possible that these clusters possess parallel functional pathways. In order to test this conceptual model, we show that concomitant lesion of A1 and A2 neuronal groups abolished compensatory mechanism, HSS-induced MAP restoration and renal sympathoinhibition after HH.
For the first time, we showed that simultaneous lesion of A1 and A2 neurons or MnPO inhibition attenuated HSS-induced MAP restoration and renal sympathoinhibition after HH. These data show the involvement of CNS components in the regulation of these reflex responses. Thus, the integrity of MnPO, A1 and A2 neuronal cluster seems to be important to the communication between peripheral afferents and central structures.

Conclusion
Taken together, our findings strengthen the hypothesis that the MnPO, A1 and A2 are part of the pathways that integrate and transmit information regarding changes in plasma osmolarity as well as modulating HSS-induced cardiovascular and autonomic responses after hypovolemia. The dysfunctions of MnPO, A1 and A2 could impair HSS-induced cardiovascular recovery after HH.