Corticotropin releasing factor excites neurons of posterior hypothalamic nucleus to produce tachycardia in rats

Corticotropin releasing factor (CRF), a peptide hormone involved in the stress response, holds a key position in cardiovascular regulation. Here, we report that the central effect of CRF on cardiovascular activities is mediated by the posterior hypothalamic nucleus (PH), an important structure responsible for stress-induced cardiovascular changes. Our present results demonstrate that CRF directly excites PH neurons via two CRF receptors, CRFR1 and CRFR2, and consequently increases heart rate (HR) rather than the mean arterial pressure (MAP) and renal sympathetic nerve activity (RSNA). Bilateral vagotomy does not influence the tachycardia response to microinjection of CRF into the PH, while β adrenergic receptor antagonist propranolol almost totally abolishes the tachycardia. Furthermore, microinjecting CRF into the PH primarily increases neuronal activity of the rostral ventrolateral medulla (RVLM) and rostral ventromedial medulla (RVMM), but does not influence that of the dorsal motor nucleus of the vagus nerve (DMNV). These findings suggest that the PH is a critical target for central CRF system in regulation of cardiac activity and the PH-RVLM/RVMM-cardiac sympathetic nerve pathways, rather than PH-DMNV-vagus pathway, may contribute to the CRF-induced tachycardia.


Results
Expression of CRFR1 and CRFR2 in the PH. We assessed the expressions of CRFR1 and CRFR2 mRNAs in the PH in rats (n = 5) by quantitative real-time RT-PCR. The rat PH tissues, localized between − 3.36 and − 4.36 mm from bregma (Fig. 1A), were extracted according to the rat brain atlas 25 . As shown in Fig. 1B, the crhr1 mRNA % v.s. gapdh mRNA was 0.3241 ± 0.0073, whereas the crhr2 mRNA % v.s. gapdh mRNA was 0.1136 ± 0.0038. Thus, both CRFR1 and CRFR2 mRNAs were detected in the rat PH.
To further map the distributions of CRFR1 and CRFR2 in the PH, we performed immunofluorescence of the rat brain slices (n = 5) containing the PH with an antibody against CRFR1 and CRFR2, respectively. We found that both CRFR1 and CRFR2 were not only localized in the PH (Fig. 1C1-C3 and D1-D3), but also co-localized in the same PH neurons (Fig. 1E1, E2 and E3).
Given that CRF exerts its physiological action via two distinct receptor subtypes 15 , CRFR1 and CRFR2, we used selective receptor agonists and antagonists to examine which CRF receptor/receptors mediated the CRF-induced excitation on PH neurons. As shown in Fig. 3, stressin I (1 μ M) and urocortin II (1 μ M), highly selective agonists for CRFR1 and CRFR2, respectively, mimicked the 300 nM CRF-induced inward current (58.7 ± 5.2 pA and 44.6 ± 4.1 pA, respectively; n = 8) on PH neurons. On the other hand, separate application of antalarmin hydrochloride and antisavagine-30, highly selective antagonists for CRFR1 and CRFR2, respectively, partly blocked the CRF-induced inward current from 70.2 ± 6.5 pA to 22.5 ± 2.1 pA and 26.7 ± 3.1 pA, respectively (n = 7, P < 0.01). Furthermore, combined application of the two antagonists nearly totally blocked the inward current (2.1 ± 0.4 pA; n = 6, P < 0.01). These results strongly suggest that both CRFR1 and CRFR2 are involved in the CRF-induced postsynaptic excitation on PH neurons.
Microinjection of CRF into the PH significantly increases HR rather than MAP via the activation of CRFR1 and CRFR2. Next, whether the excitatory action of CRF on PH neurons influences cardiovascular activities in vivo was assessed. We found that microinjection of 20, 60 and 200 μ M CRF into the rat PH elicited a remarkable increase in average peak HR (6.3 ± 0.8, 13.8 ± 1.7 and 25.8 ± 3.1 bpm, respectively) in a concentration-dependent manner, compared to no change (3.2 ± 0.7 bpm) following injection of normal saline (n = 8, P < 0.01; Fig. 4A1,4A2). However, microinjecting CRF (up to 200 μ M) into the PH did not influence MAP ( Fig. 4B1 and B2) and RSNA ( Fig. 4C1 and C2). The changes in MAP after injection of CRF or vehicle were 2.8 ± 0.8 mmHg and 1.7 ± 0.8 mmHg (n = 8, P = 0.36), respectively. And the change ratios of integrated RSNA were 6.3 ± 2.9% v.s. 6.1 ± 2.4% (n = 8, P = 0.16). The data indicate that the PH contributes to potentiation effect of CRF on HR rather than AP and RSNA.
Given an involvement of both CRF receptors in the excitation of CRF on PH neurons, we further addressed whether CRFR1 and CRFR2 in the PH co-mediate the enhancement effect of CRF on HR. In this experiment, selective CRFR1 antagonist antalarmin hydrochloride or selective CRFR2 antagonist antisauvagine-30 was microinjected into the PH, followed within 2 min by a microinjection of CRF at the same site. As shown in Fig. 5, blockade of either CRFR1 or CRFR2 in the PH partly reduced the tachycardic effect induced by CRF. The increase in average peak HR induced by CRF decreased from 25.8 ± 3.1 bpm to 7.8 ± 2.1 bpm (n = 5, P < 0.05) and 13.8 ± 1.9 bpm (n = 5, P < 0.05), respectively. Moreover, microinjection of broad-spectrum CRF receptor antagonist astressin almost totally blocked the tachycardic response to microinjection of CRF into the PH (the increase in average peak HR decreased to 2.6 ± 1.3 bpm; n = 5, P < 0.01), suggesting that both CRFR1 and CRFR2 in the PH are involved in the tachycardic effect of CRF.

Effect of vagotomy and blockade of CSNA on the CRF-induced tachycardia.
To clarify the peripheral mechanism underlying the tachycardia effect of CRF mediated by the PH, we determined whether the vagus nerve or cardiac sympathetic nerve was involved in. As shown in Fig. 6, bilateral vagotomy increased the baseline of HR; however, it only slightly reduced the tachycardic response elicited by microinjection of CRF into the PH. The increases in HR induced by microinjection of CRF before and after the vagotomy were 25.8 ± 3.1 bpm and 22.0 ± 2.1 bpm, respectively (n = 5, P = 0.56). On the other hand, blocking CSNA by intravenous injection of propranolol almost abolished the tachycardic response elicited by microinjection of CRF into the PH. The increase in HR induced by microinjection of CRF after the blockade of CSNA reduced to 1.4 ± 1.0 bpm (n = 5, P < 0.01; Fig. 6). These results indicate that it is the cardiac sympathetic nerve rather than vagus mediates the tachycardia effect induced by microinjection of CRF into the PH. Responses of RVLM, RVMM, and DMNV neurons to microinjection of CRF into the PH. Considering that the RVLM and RVMM in the brainstem are two most important areas of presympathetic neurons, whereas the DMNV and nucleus ambiguus are two nuclei gathering vagal preganglionic neurons 28 , we conducted extracellular recordings in the RVLM, RVMM and DMNV (the nucleus ambiguus neurons were not recorded since they are intrinsically silent 29 ) in vivo to clarify which structures were involved in the tachycardic effect elicited by microinjection of CRF into the PH. We recorded 80 RVLM, 72 RVMM and DMNV neurons in rats (n = 12). Among them, 37 (37/80, 46.3%) RVLM neurons and 32 (32/72, 44.4%) RVMM neurons increased their firing rates from 9.3 ± 1.2 spikes/s and 12.8 ± 1.9 spikes/s to 15.7 ± 1.8 spikes/s (P < 0.01) and 21.3 ± 2.5 spikes/s (P < 0.01), respectively, in response to microinjection of CRF into the PH (Fig. 7). In addition, 4 (4/80, 5.0%) RVLM neurons and 3 (3/72, 4.2%) RVMM neurons decreased their firing rates from 21.0 ± 2.8 spikes/s and 23.8 ± 3.8 spikes/s to 12.3 ± 2.0 spikes/s (P < 0.05) and 12.8 ± 2.4 spikes/s (P < 0.05), respectively, to microinjection of CRF into the PH (Fig. 7), and the rest 39 RVLM neurons and 37 RVMM neurons showed no response. On the other hand, all of the 14 recorded DMNV neurons had no response (1.9 ± 0.3 spikes/s v.s. 1.7 ± 0.2 spikes/s, P = 0.58) to microinjection of CRF into the PH (Fig. 7). The results suggest that it is the RVLM and RVMM but not DMNV mediate the tachycardia effect induced by microinjection of CRF into the PH.

Discussion
CRF holds a key position in regulating and integrating various stress responses 1,2 and the linkage between stress and cardiovascular disease risk has attracted a growing attention 3,4,30 . However, mechanism of central regulation of CRF on cardiovascular activities remains enigmatic. Here, we demonstrate a novel contribution of the PH, a pivotal structure responsible for generation of stress-induced cardiovascular responses, in the central effect of CRF on cardiovascular activities. CRFR1 and CRFR2 are expressed and co-localized in the same neurons in PH in rats. Through activation of both CRFR1 and CRFR2, CRF directly excites the PH neurons and actively elevates HR rather than AP. Furthermore, the tachycardia effect induced by microinjecting CRF into the PH is mediated by PH-RVLM/RVMM-cardiac sympathetic nerve pathway rather than PH-DMNV-vagus nerve pathway.
CRFR1 and/or CRFR2 are found to be expressed in many areas of cardiovascular centers including the hypothalamic paraventricular nucleus, arcuate nucleus, and nucleus tractus solitarii, nucleus ambiguus, and RVLM in brainstem [31][32][33][34][35][36][37] . Via coupling to various ionic mechanisms, CRFR1 and CRFR2 may mediate different effect of CRF related peptides on these neurons. For instance, CRFR1 and its coupled hyperpolarization-activated cation channel mediate the CRF-induced excitation 38 , whereas CRFR2 and its linked G protein-activated inwardly rectifying potassium channel mediate the urocortin 3-evoked inhibition on paraventricular nucleus neurons 39 . By these ways, both of CRF receptors may actively participate in central modulation of cardiovascular activities [31][32][33][34][35][36][37] . In the present study, we report, for the first time, that CRFR1 and CRFR2 are co-expressed in the PH and co-mediate the CRF-induced excitation on PH neurons (Figs 1-3). Since either excitation by electrical stimulation or disinhibition by microinjection GABA receptor antagonist in the PH significantly increased AP and HR 19,40 , our result that CRF excited the PH neurons suggests a possibility that PH may contribute to central regulation of CRF on cardiovascular activities. Intriguingly, microinjection of CRF into the PH significantly increased HR but not MAP and RSNA (Fig. 4), suggesting that by activating PH neurons, CRF may actively affect cardiac activity. Furthermore, selective CRFR1 antagonist antalarmin or selective CRFR2 antagonist antisauvagine-30 partly decreased tachycardiac response to microinjecting CRF into the PH, while broad-spectrum CRF receptor antagonist astressin totally blocked the response (Fig. 5), demonstrating that the tachycardiac response to CRF microinjected into the PH is mediated by activation of both CRFR1 and CRFR2. Nonetheless, microinjection of selective antagonist for CRFR1 or CRFR2 or broad-spectrum CRF receptor antagonist into the PH before CRF injection did not elicit any changes in basal HR (Fig. 5A). Thus, we speculate that under normal physiological conditions, CRF receptors in the PH may be not under a status of tonic control of endogenous CRF to modulate cardiovascular activity.
Microinjection of CRF into the PH only elevated HR but did not influence AP and RSNA suggests that the regulation of cardiac activity by CRF may be mediated by autonomic cardiovascular centers. On the basis of current knowledge regarding the pathway underlying the PH control of cardiac function, activation of cardiac sympathetic nerve or inhibition of cardiac vagal nerve may be responsible for tachycardiac effect followed by microinjection of CRF into the PH. In this study, we found that vagotomy did not affect the tachycardiac response by CRF, while blockade of CSNA with propranolol almost totally abolished the tachycardiac response (Fig. 6). These results reveal that the tachycardia response is not an effect of reduced vagus nerve activity elicited by microinjecting CRF into PH, but due to an increase in CSNA. This result accords with the previous reports that activation of the PH increased HR by increasing sympathetic nerve activity 41,42 . It is well known that the PH  neurons project to various areas/structures in the brainstem, which are involved in regulation of cardiovascular activity, including the RVLM and RVMM 43 . The RVLM and RVMM are the key centers providing synaptic inputs to sympathetic preganglionic neurons and directly regulating sympathetic activity to influence cardiovascular function. Excitation of RVLM 44 and RVMM 45 increases HR and AP by strengthening sympathetic nerve activity which output through the sympathetic preganglionic neurons of intermediolateral cell column in spinal cord. Thus, the tachycardiac effect induced by microinjection of CRF in the PH may be mediated by the excitation of RVLM and RVMM neurons. In fact, in this study, we found that it is the RVLM and RVMM neurons rather than DNV neurons controlling vagus activity responsive to microinjecting CRF into the PH. Therefore, we speculate that the pathways from the PH to the RVLM and RVMM may contribute to the modulation of central CRF system on sympathetic outputs to the heart and the consequent HR regulation.
Tachycardia is an important component and feature of stress-induced cardiovascular responses and dysfunctions. We demonstrate that CRF, a critical neuropeptide responsible for initiating and integrating stress responses, induces a remarkable tachycardia by exciting PH neurons, which has been proposed to participate in the generation of stress-induced tachycardia 23 , through both CRFR1 and CRFR2. Pathways from the PH to the RVLM and RVMM, which subsequently regulate sympathetic outflow, may contribute to the tachycardia induced by activation of PH by CRF. The findings not only help to understand the effect and mechanism of central CRF on cardiovascular regulation, but also provide a new insight into the stress-related cardiovascular dysfunctions and disorders.

Methods
Animals. We used male Sprague-Dawley rats (375 ± 50 g), housed under controlled conditions with a lighting schedule of 12 h light and 12 h dark at a temperature of 22 ± 2 °C. Standard food and water were provided ad libitum. All experiments, approved by the Experimental Animal Care and Use Committee of Nanjing University, were treated in accordance with U.S. National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publication 85-23, revised 2011) and were reported in accordance with the ARRIVE guidelines 46 . All efforts were made to minimize the number of animals used and their suffering.
Quantitative real-time RT-PCR. For this experiment, three independent groups of RNA pools each from 5 rats, deeply anaesthetized with urethane (1600 mg/kg) plus α -chloralose (130 mg/kg) given intraperitoneally, were used as biological replicates. The PH tissue punches were collected from coronal brain slices of rats according to the rat brain atlas of Paxinos and Watson 25 and pooled (5 animals in each pool). RNA extraction was done using an RNA isolation kit (Invitrogen, USA) according to the manufacturer's instructions. The reaction was carried out in an Applied Biosystem 7300 Real-time PCR System using the following parameters: 47 cycles at 95 °C

Figure 7. Responses of RVLM, RVMM and DMNV neurons to microinjection of CRF into the PH in rats.
(A1,A2) The upper and middle peristimulus time histograms of A1 show two single RVLM neurons increased or decreased firing rate in response to microinjection of CRF into the PH, respectively. The lower histogram in A1 shows a non-responsive RVLM neuron to the microinjection. The group data with statistical results obtained from 80 recorded RVLM neurons is shown in A2. Note that 46.3% (37/80) RVLM neurons were excited. (B1,B2) RVMM neurons also showed increase, decrease, or no change in firing rate in response to microinjection of CRF into the PH, respectively. The layout of this panel is the same with (A1,A2). (C1,C2) Firing rate of DMNV neurons exhibited no change in response to microinjection of CRF into the PH. Statistical significance was determined using Student's t test. n.s. no significance, *P < 0.05, **P < 0.01.
Scientific RepoRts | 6:20206 | DOI: 10.1038/srep20206 for 5 s, 60 °C for 30 s, 95 °C for 15 s, 55 °C for 30 s and 95 °C for 15 s. The PCR program was completed by a melting temperature analysis. For quantification, the quantity of the target gene (crfr1and crfr2) was expressed relative to the amount of the reference gene (gapdh) to obtain a normalized target expression value.
Briefly, rats (n = 5) were deeply anesthetized and the rat brain was removed and fixed in the 4% paraformaldehyde in 0.1 M phosphate buffer for 12 h at 4 °C, and then cryoprotected with 30% sucrose for 48 h. Frozen coronal sections (25 μ m thickness) containing the PH were obtained by using a freezing microtome (CM 1850, Leica, Germany). The slices were rinsed in phosphate-buffered saline containing 0.1% Triton X-100 (PBST) (Sigma, USA) and then incubated in 10% normal bovine serum (Millipore, USA) in PBST for 30 min. Sections were incubated overnight at 4 °C with primary antibodies to CRFR1 (a rabbit anti-CRFR1 polyclonal antibody, 1:200; Abcam, USA) and CRFR2 (a goat anti-CRFR2 polyclonal antibody, 1:200; Abcam). After a complete wash in PBS, sections were incubated in the Alexa 594-conjugated donkey anti-rabbit (1:2,000; Invitrogen) and Alexa 488-conjugated donkey anti-goat (1:2,000; Invitrogen) for 2 h at room temperature in the dark. The slides were washed and mounted in UltraCruz mounting medium (Santa Cruz Biotechnology, USA). All micrographs were taken with an inverted laser scanning confocal microscope (FV1000; Olympus, Japan).
Patch clamp recordings on brain slices. Rats aged 14-21 days were deeply anaesthetized and coronal brainstem slices (300 μ m in thickness) containing the PH were prepared with a vibroslicer (VT 1200 S, Leica), according to the rat brain atlas 25 . The slices were incubated in artificial cerebrospinal fluid (ACSF, composition in mM: 124 NaCl, 2.5 KCl, 1.25 NaH 2 PO 4 , 1.3 MgSO 4 , 26 NaHCO 3 , 2 CaCl 2 and 10 D-glucose) equilibrated with 95% O 2 and 5% CO 2 at 35 ± 0.5 °C for at least 1 hour and then maintained at room temperature. During recording sessions, the slices were transferred to a submerged chamber and continuously superfused with 95% O 2 and 5% CO 2 oxygenated ACSF at a rate of 2 ml/min maintained at room temperature.
Whole-cell recordings were performed as we previously described 47,48 . Briefly, PH neurons were visualized with an Olympus BX51WI microscope and then recorded with borosilicate glass pipettes (3)(4)(5)  Cardiovascular activities recording in vivo. Rats were anaesthetized with urethane (800 mg/kg) plus α -chloralose (65 mg/kg) given intraperitoneally and maintained at surgical depth throughout the experimental session. Supplemental doses of urethane plus α -chloralose were given intravenously when necessary if nociceptive stimulation of hind paw (tested every half an hour) caused a change in MAP of more than 10 mmHg. Rectal temperature was maintained at 37 ± 0.5 °C. The trachea was intubated for artificial respiration. The right carotid artery was cannulated with polyethylene tubing (i.d. 0.5 mm, o.d. 0.9 mm) filled with normal saline containing heparin (30 U/ml, Sigma, USA) for the measurement of AP. The polyethylene tubing was connected to a physiological pressure amplifier (Bridge Amp, ADInstruments, Australia) through a transducer (SP844, ADInstruments). MAP and HR were derived from the AP tracing. All signals were acquired online using LabChart 7 software (ADInstruments). The left renal sympathetic nerve was isolated and cut distally to permit recording of efferent RSNA. Nerve recordings were made with bipolar silver wire electrodes. The renal nerve branch and the electrode were immersed in warm liquid paraffin. The signals from the recording electrodes were amplified (× 1000) and filtered (bandwidth 100-2,000 Hz). These signals were recorded on a computer using a PowerLab system (ADInstruments). LabChart 7 software was used to rectify and integrate the renal sympathetic nerve signals.
Bilateral cervical vagus nerves were isolated and tied with silk threads. The vagus nerves were cut by carefully pulling on the silk loop to detect the CRF-induced changes in HR after bilateral vagotomy followed a stabilization period of 15-20 min. To assess role of cardiac sympathetic nerve in the CRF-induced changes in HR, the jugular vein was cannulated for administrating propranolol (1 mg/kg, Sigma) to block the cardiac sympathetic nerve activity (CSNA).
Microinjection in the PH. The rat was mounted on a stereotaxic frame (1404, David Kopf Instruments, USA). The craniotomy was made above the PH. According to the rat brain atlas 25  , and saline (0.9% NaCl) using Hamilton syringes (0.5 μ l, lasting 2 min). The effective extent of the drug diffusion in the present study was restricted in the PH according to the estimate by using extracellular electrophysiological recording neurons 0.5 mm away from the injection site (the detailed descriptions of methodology can be obtained from our previous reports 47,48 ). The effects of CRF on HR, MAP and RSNA before and after microinjections were detected.
Scientific RepoRts | 6:20206 | DOI: 10.1038/srep20206 Extracellular recording in brainstem in vivo. Single neuronal discharges were recorded from the RVLM (A − 12.00 to − 12.48, L 1.8 to 2.4, H 8.0 to 8.5), RVMM (A − 10.00 to − 12.20, L 0.0 to 0.6, H 8.5 to 9.5), and DMNV (A − 13.08 to − 13.20, L 0.5 to 1.0, H 5.2 to 5.5), according to the rat brain atlas 25 , by using a glass electrode filled with 1% solution of pontamine sky blue in 0.5 M sodium acetate (DC resistance 5-10 MΩ ). The neuronal discharges were conventionally amplified, displayed on an oscilloscope, and simultaneously fed into a window discriminator. The standard rectangular pulses (1.0 ms, 5 V) triggered from the spikes were sent through an A/D interface (Power 1401, CED, UK) into a laboratory computer.
The peristimulus time histogram (sampling interval 1 sec, sampling length 1,800 sec, single trial) of neuronal discharges were generated by the software Spike 2 (CED) to assess the effects of microinjection of CRF into PH on neuronal discharges of RVLM, RVMM, and DMNV, respectively. According to the previous established criteria 49,50 , any significant change in neuronal discharge rate, reproducibly evoked by the microinjection, with P < 0.05 was considered as an excitatory or inhibitory response. After experiments, microinjection sites in the PH and recording sites in the RVLM, RVMM, and DMNV were histologically identified as we previously reported [47][48][49][50] . Data from rats in which the injection and the recording sites were out of the target nucleus were excluded from further analysis. Statistical analysis. All data are presented as mean ± SEM. Preliminary tests for normality of each variable were executed. The Student's t test was employed for statistical analysis. P < 0.05 was considered significant.