Endogenous hydrogen peroxide in the hypothalamic paraventricular nucleus regulates sympathetic nerve activity responses to l-glutamate

Abstract

The hypothalamic paraventricular nucleus (PVN) is important for maintenance of sympathetic nerve activity (SNA) and cardiovascular function. PVN-mediated increases of SNA often involve the excitatory amino acid l-glutamate (l-glu), whose actions can be positively and negatively modulated by a variety of factors, including reactive oxygen species. Here, we determined modulatory effects of the highly diffusible reactive oxygen species hydrogen peroxide (H₂O₂) on responses to PVN l-glu. Renal SNA (RSNA), arterial blood pressure, and heart rate were recorded in anesthetized rats. l-Glu (0.2 nmol in 100 nl) microinjected unilaterally into PVN increased RSNA (P < 0.05), without affecting mean arterial blood pressure or heart rate. Effects of endogenously generated H₂O₂ were determined by comparing responses to PVN l-glu before and after PVN injection of the catalase inhibitor 3-amino-1,2,4-triazole (ATZ; 100 nmol/200 nl, n = 5). ATZ alone was without effect on recorded variables, but attenuated the increase of RSNA elicited by PVN l-glu (P < 0.05). PVN injection of exogenous H₂O₂ (5 nmol in 100 nl, n = 4) and vehicle (artificial cerebrospinal fluid) were without affect, but H₂O₂, like ATZ, attenuated the increase of RSNA to PVN l-glu (P < 0.05). Tonic effects of endogenous H₂O₂ were determined by PVN injection of polyethylene glycol-catalase (1.0 IU in 200 nl, n = 5). Whereas polyethylene glycol-catalase alone was without effect, increases of RSNA to subsequent PVN injection of l-glu were increased (P < 0.05). From these data, we conclude that PVN H₂O₂ tonically, but submaximally, suppresses RSNA responses to l-glu, supporting the idea that a change of H₂O₂ availability within PVN could influence SNA regulation under physiological and/or disease conditions.

presympathetic neurons of the hypothalamic paraventricular nucleus (PVN) contribute to generation of sympathetic nervous system and cardiovascular responses to physiological challenges, including body fluid hyperosmolality/water deprivation (8, 14, 20, 26, 43), hemorrhage (6), hyperthermia (12), and hyperinsulinemia (49). They also participate in chronic cardiovascular and metabolic diseases, such as arterial hypertension (18, 35), heart failure (27, 33), obesity (39), and diabetes (21, 49).

Although neurochemical mechanisms of PVN-driven sympathetic and cardiovascular responses are varied and complex, neuronal activation by the excitatory amino acid l-glutamate (l-glu) is an important factor in many cases (20, 27, 29, 31–33). During water deprivation, for example, presympathetic PVN neurons are recruited (43–45), at least in part, through activation of ionotropic glutamate receptors (20), and this is vital for support of sympathetic nerve activity (SNA), arterial blood pressure (ABP), and heart rate (HR) (20, 43–45).

Glutamatergic activation of PVN is also required for increases of SNA, ABP, and HR evoked by pharmacological blockade of local GABA_A receptors (13, 34). Functionally, reduced GABAergic inhibition of PVN contributes to sympathetic hyperactivity in models of cardiovascular disease, including chronic heart failure (55) and renal wrap hypertension (23, 35). In the case of heart failure, reduced GABAergic inhibition is accompanied by 1) hyperactivity of presympathetic PVN neurons (50, 56); 2) increased expression of PVN N-methyl-D-aspartate (NMDA) receptors (33); and 3) enhanced PVN glutamatergic support of SNA, ABP, and HR (33, 40).

Because glutamatergic transmission in PVN plays such a prominent role in controlling SNA and cardiovascular function, factors that modulate it are important to fully understand. Reactive oxygen species (ROS) are of particular interest and importance in this regard. Not only are various ROS family members potent modulators of glutamatergic transmission (2, 3, 5, 48, 61), but ROS activity plays a key regulatory role in regions of brain that control sympathetic drive (10, 22, 28, 53, 57, 60).

In the present study, we focused on the role of hydrogen peroxide (H₂O₂), which is mainly generated from its precursor superoxide (O₂^·−) by superoxide dismutase (SOD) (41). H₂O₂ has several known neuromodulatory effects (5, 41) and can inhibit glutamatergic neurotransmission, both by reducing calcium-dependent l-glu release (61) and by directly inhibiting ionotropic l-glu receptors by acting at their redox modulatory sites (1–3). H₂O₂ can also enhance glutamatergic transmission by reducing astrocytic sodium-dependent l-glu uptake (48), with the latter action being consistent with effects of H₂O₂ in the nucleus tractus solitarius (10), a key brain stem region involved in reflex control of SNA (36).

Because H₂O₂ is more stable and diffusible than other ROS, including O₂^·− (41), we reasoned that experimentally induced changes of H₂O₂ concentration in PVN could have wide spatial impact and could, therefore, significantly modulate SNA responses to stimulation with l-glu. Because tissue H₂O₂ concentration is largely determined by the balance of enzymatic production (i.e., from O₂^·− by SOD) and elimination (i.e., “scavenging” by catalase and glutathione peroxidase) (41, 60), we adopted the following experimental strategy. First, we determined effects of inhibiting catalase activity in PVN on renal SNA (RSNA), ABP, and HR responses to locally delivered l-glu. Next, we assessed whether effects of catalase inhibition were attributable to an increase of tissue H₂O₂ by determining if PVN microinjection of H₂O₂ mimicked effects of local catalase inhibition. Finally, we tested for tonic modulation by endogenous H₂O₂ by determining effects of PVN microinjection of catalase on l-glu-evoked responses. Our findings indicate 1) that endogenous H₂O₂ in PVN tonically suppresses l-glu-evoked increases of RSNA; and 2) that tonic suppression is submaximal, since increasing PVN H₂O₂ further attenuates RSNA responses to locally delivered l-glu.

MATERIALS AND METHODS

Animals

Adult male Sprague-Dawley rats (Charles River Laboratories), weighing 350–450 g, were housed in a temperature-controlled (22–23°C) room with a 14:10-h light-dark cycle (lights on at 0700). Tap water and standard chow (Harlan Teklad LM-485) were available ad libitum. All protocols and surgical procedures were approved by the Institutional Animal Care and Use Committee of the University of Texas Health Science Center at San Antonio.

Chemicals

3-Amino-1,2,4-triazole (ATZ), H₂O₂, polyethylene glycol-catalase (PEG-catalase, 35,663 U/mg protein, 54 mol of PEG/mol protein), gallamine triethiodide, hexamethonium, urethane, and α-chloralose were obtained from Sigma (St. Louis, MO). Gallamine triethiodide was dissolved in a 1.5% glucose solution. All other drugs were dissolved in phosphate-buffered saline (PBS) with pH adjusted to 7.30.

Animal Preparation and Surgery

Rats were anesthetized with isoflurane (3% in oxygen) and catheters (PE-10 fused to PE-50 tubing, Clay Adams, Parsippany, NJ) were placed in a femoral artery and vein to monitor ABP and inject drugs, respectively. HR was obtained from the R-R interval of a lead I ECG. Rectal temperature was continuously monitored (YSI, Yellow Spring, OH) and maintained at 37 ± 1°C with a water-circulating pad (Baxter Healthcare, Deerfield, IL). Anesthesia was gradually switched to a mixture of α-chloralose (70 mg/kg iv) and urethane (700 mg/kg iv), a cannula was placed in the trachea, and rats were artificially ventilated (Kent Scientific) with 100% O₂-enriched room air. End-tidal Pco₂ was continuously monitored (CapStar 100, CWE) and maintained between 40 and 45 Torr by adjusting ventilation rate (80–90 breaths/min) and/or tidal volume (2.0–2.5 ml). Rats were placed in a stereotaxic apparatus (David Kopf), paralyzed with gallamine triethiodide (5 mg/kg iv bolus, 5 mg·kg⁻¹·h⁻¹ iv infusion), and activity was recorded from a left renal nerve bundle, as previously described (14, 42, 43). Before and after paralysis, adequacy of anesthesia was assessed by lack of an ABP or HR response to noxious pinching of a hindpaw. Anesthetic supplements (5–10% of initial dose, iv) were given as needed.

PVN Microinjections

After recordings were established, a craniotomy was performed, and a three-barreled glass micropipette (tip diameter: 40–60 μm, outer diameter) was vertically lowered into PVN at the following coordinates: −2.1 mm caudal to bregma, 0.4 mm lateral to midline, and −7.5 mm ventral to the surface of cortex. The position of each micropipette was adjusted slightly to identify the site from which unilateral injection of l-glu (0.2 nmol) elicited the largest increase of RSNA. Injections were delivered using a hydraulic coupled volume microinjection system (38), which consisted of each micropipette barrel being connected to a 5-μl Hamilton syringe with PE-10 tubing filled with light mineral oil so that no air remained in the system. The volume ejected from each micropipette is determined by the plunger displacement of the attached syringe, which was controlled by a programmable micropump (Micro4, WPI, Sarasota, FL). This system provides consistent control of the volume and time course of injections, allowing slow and continuous infusion of solution into the PVN (38). Depending on the experiment, volumes injected were 100 or 200 nl delivered at rates of 1 to 5 nl/s (see below).

Protocols

Inhibition of catalase.

The catalase inhibitor ATZ was microinjected unilaterally into PVN to investigate effects of reduced H₂O₂ scavenging on RSNA, ABP, and HR responses to local microinjection of l-glu (n = 5). Two consistent RSNA responses to l-glu (0.2 nmol in 100 nl, 5 nl/s) were obtained, and 5–10 min later ATZ (100 nmol in 100 nl, 2 nl/s) was microinjected in the same site. Microinjection of l-glu was repeated 1–2 min thereafter to test for effects of catalase inhibition. To test for recovery, l-glu was injected again 45 min later.

Exogenous H₂O₂.

Inhibition of H₂O₂ scavenging by catalase is expected to increase endogenous levels of H₂O₂ in PVN. To verify that effects of ATZ could be attributed to actions of elevated H₂O₂, we next tested direct effects of exogenous H₂O₂ on responses to PVN injection of l-glu (n = 5). As described above, two consistent RSNA responses to l-glu were obtained, followed by injection of H₂O₂ (5 nmol in 100 nl) at the same site 5–10 min later. H₂O₂ was delivered at a rate of 1.0 nl/s, which was determined in preliminary experiments to have no effect on resting RSNA, thereby mimicking actions of ATZ. Microinjections of l-glu were performed 1–2 and 45 min after injection of H₂O₂.

Exogenous catalase.

To determine the impact of an acute increase of H₂O₂ scavenging in PVN on responses to microinjection of l-glu, membrane-permeable PEG-catalase was used. Rats (n = 10) were divided into two equal groups: one received PVN injection of PEG-catalase, and the other an equal molar concentration of vehicle PEG (4,000 mol wt). As described above, 5–10 min after recording two consistent RSNA responses to PVN l-glu, either PEG-catalase (1.0 IU) or PEG vehicle (200 nl, 2 nl/s) was microinjected into the same site. l-Glu injections were repeated 1–2 and 45 min later.

Histology

After completing each microinjection experiment, 100 nl of 2% Chicago Sky Blue dye in isotonic saline was microinjected unilaterally into the PVN at the same coordinates used for injection of l-glu. Each rat was then decapitated and the brain was removed, postfixed in 4% paraformaldehyde (in PBS) for 24 h, and sectioned at 50 μm using a freezing microtome. Dye distribution for each rat was determined as previously described (42, 43). Briefly, digital images of PVN sections were taken using a charge-coupled device camera (Sony), and the outermost dye boundary was traced. Traced areas were then mapped onto appropriate rostro-caudal plates of PVN. Traced outlines from similar rostra-caudal levels of PVN were overlaid, and a final trace was made of the largest distribution of dye for each group.

Data Acquisition and Analysis

Recorded signals (RSNA, ABP, ECG, body temperature, and end-tidal Pco₂) were digitized and stored on a computer hard disk for offline analysis using Spike2 software (version 6, Cambridge Electronic Design, Cambridge, UK). Pulsatile ABP was smoothed with an RC filter (time constant = 0.5 s) to obtain mean arterial pressure (MAP). HR was determined as the mean frequency of R-R intervals obtained from ECG recordings. RSNA was determined from the rectified and integrated signal after subtracting noise, which was determined as the signal remaining 5 min after ganglionic blockade with hexamethonium (30 mg/kg iv). Baseline RSNA was set at 100%, and responses are reported as a percentage of baseline. Responses of RSNA, MAP, and HR were determined as the difference between a 30-s baseline average and the maximal value of each variable recorded within 60 s of each PVN injection.

Two-way repeated-measures ANOVA was used to compare time-dependent effects of PEG-catalase and PEG vehicle on l-glu-evoked responses and to compare effects of ATZ, H₂O₂, PEG-catalase, and PEG vehicle on resting RSNA, MAP, and HR. Effects of ATZ and H₂O₂ on l-glu-evoked responses were compared separately, each with one-way repeated-measures ANOVA. Pairwise comparisons were made using Bonferroni post hoc tests. P < 0.05 was considered significant. Data in the text and Figs. 2 and 4 are reported as means ± SE.

Fig. 2.

Summary of effects of ATZ and H₂O₂ on responses to PVN l-glu. Unilateral PVN injection of l-glu was without effect on HR (A) and mean arterial pressure (MAP; B) in all groups, but significantly increased RSNA (C) under control conditions, with increases being significantly smaller when elicited immediately (1–2 min) after PVN injection of ATZ (open bars) and H₂O₂ (solid bars). RSNA responses to l-glu recovered toward control 45 min later. Δ, Change. Values are means ± SE. *P < 0.05, peak value compared with baseline. †P < 0.05, compared with corresponding group response under control conditions and to vehicle treatment at the same time point.

Fig. 4.

Summary of effects of catalase on responses to PVN l-glu. Unilateral PVN microinjection of l-glu was without effect on HR (A) or MAP (B) in any treatment group. C: however, PVN l-glu significantly increased RSNA under control conditions, with increases being significantly greater when elicited immediately (1–2 min) after PVN injection of PEG-cat (open bars). Recovery was observed within 45 min. PVN injection of PEG vehicle (solid bars) had no effect on responses to PVN l-glu. Values are means ± SE. *P < 0.05, compared with control and 45-min time points and PEG vehicle at the same time point.

RESULTS

Baseline Values

Baseline values of RSNA, MAP, and HR are given in Table 1. Values in vehicle (PEG or PBS) and treatment (ATZ, H₂O₂, PEG-catalase) groups were not significantly different.

Table 1.

Baseline values of recorded variables in each treatment group

Treatment	n	RSNA, %	MAP, mmHg	HR, beats/min
ATZ	5	100	105 ± 4	385 ± 16
H2O2	4	100	105 ± 3	370 ± 11
PEG-catalase	5	100	109 ± 2	386 ± 13
PEG	4	100	96 ± 9	412 ± 23
PBS	4	100	107 ± 5	382 ± 11

Values are means ± SE; n, no. of rats. RSNA; renal sympathetic nerve activity, MAP; mean arterial pressure, HR; heart rate; ATZ, 3-amino-1,2,4-triazole; H₂O₂, hydrogen peroxide; PEG, polyethylene glycol; PBS, phosphate-buffered saline. Values of each variable were compared by one-way ANOVA. No significant differences were detected.

Effects of Catalase Inhibition and Exogenous H₂O₂ on Responses to PVN l-glu

Representative control responses to unilateral PVN microinjection of l-glu are shown in Fig. 1, left. Whereas ABP and HR were unaffected, RSNA transiently (30–120 s) increased. Immediately (1–2 min) after injection of the catalase inhibitor ATZ at the same site (Fig. 1A, middle), the increase of RSNA evoked by PVN l-glu was noticeably attenuated, but returned toward control after 45 min (Fig. 1A, right). Group data (n = 5) are summarized in Fig. 2 (open bars) and reveal that, whereas HR (Fig. 2A) and MAP (Fig. 2B) responses to PVN l-glu were similar under control conditions and immediately after ATZ, the average increase of RSNA was significantly (P < 0.05) reduced (12 ± 1%) compared with control (23 ± 2%). The average RSNA response to PVN l-glu recovered nearly to control 45 min after injection of ATZ.

Fig. 1.

Effects of local 3-amino-1,2,4-triazole (ATZ; A) and hydrogen peroxide (H₂O₂; B) on responses to paraventricular nucleus (PVN) l-glutamate (l-glu). Under control conditions (left), PVN injection of l-glu (0.2 nmol in 100 nl) increased renal sympathetic nerve activity (RSNA), without affecting arterial blood pressure (ABP) or heart rate (HR). PVN microinjection of the catalase inhibitor ATZ (100 nmol in 100 nl; A, middle) or H₂O₂ (5 nmol in 100 nl; B, middle) each reduced the increase of RSNA elicited by PVN injection of l-glu immediately (1–2 min) thereafter. C, middle: vehicle injection had no effect. Note: the start of each injection is indicated by an arrowhead (top) and vertical line in each panel. Injection duration is indicated by a thick line at bottom. bpm, Beats/min; au, arbitrary units.

To assess whether the inhibitory effect of PVN ATZ was attributable to an elevation of local H₂O₂ concentration during catalase inhibition, experiments tested if similar inhibitory effects could be produced by PVN injection of exogenous H₂O₂. As Fig. 1B (middle) shows, immediately (1–2 min) after PVN injection of H₂O₂, the increase of RSNA by PVN l-glu was reduced. Group data (n = 4) in Fig. 2 (solid bars) show that the average increase of RSNA by PVN l-glu immediately (1–2 min) after injection of H₂O₂ (10 ± 2%) was significantly (P < 0.05) smaller than under control conditions (18 ± 2%). As with ATZ, RSNA responses to PVN l-glu returned toward control 45 min after injection of H₂O₂.

To ensure that blunted RSNA responses to PVN l-glu were not due to tissue damage or local edema caused by prior PVN injections of ATZ or H₂O₂, vehicle (PBS) was injected in place of ATZ/H₂O₂. Representative traces in Fig. 1C and group data in Fig. 2, A–C (hatched bars, n = 4), show that vehicle had no effect on RSNA. Note that average increases of RSNA by PVN l-glu immediately (1–2 min) after injection of ATZ (12 ± 1%) and H₂O₂ (10 ± 2%) were each significantly (P < 0.01) smaller than the corresponding increase elicited after injection of vehicle (23 ± 5%).

Effects of Catalase on Responses to PVN l-glu

Figure 3 shows representative responses to PVN l-glu before (left) and after (middle and right) PVN injections of PEG-catalase (Fig. 3A) and PEG vehicle (Fig. 3B). Whereas increases of RSNA by PVN l-glu before (Fig. 3A, left) and 45 min after PEG-catalase (Fig. 3A, right) were similar, the response immediately (1–2 min) after PEG-catalase was larger (Fig. 3A, middle). PEG vehicle had no effect on RSNA responses to PVN l-glu at any time point (Fig. 3B), suggesting that the enhanced RSNA response to PVN l-glu was likely due to H₂O₂ scavenging by catalase. Group data in Fig. 4, A and B, shows that HR and MAP were unaffected by PVN l-glu before or after PEG vehicle (solid bars) or PEG-catalase (open bars). Figure 4C shows that the increase of RSNA by PVN l-glu was unaffected by PEG vehicle (solid bars). In contrast, before injection of PEG-catalase (open bars, n = 5), PVN l-glu increased RSNA by an average of 19 ± 2%. Immediately (1–2 min) thereafter, the response was significantly increased (P > 0.05) to 34 ± 6%. l-Glu evoked RSNA responses after PEG-catalase returned toward control within 45 min.

Fig. 3.

Effects of catalase on responses to PVN l-glu. Under control conditions (left), unilateral PVN injection of l-glu (0.2 nmol in 100 nl) increased RSNA, without affecting ABP or HR. A: immediately (1–2 min) after injection of polyethylene glycol-catalase (PEG-cat; 1.0 IU in 200 nl) into PVN at the same site (middle), the increase of RSNA by PVN l-glu was larger (middle) with recovery observed after 45 min (right). B: injection of PEG vehicle did not changed the l-glu-evoked response at any time point. Note: the start of each PVN injection is indicated in each panel by an arrowhead (top) and vertical line. Injection duration is indicated by a thick line at bottom.

Histology

Targeting of l-glu microinjections to PVN was determined from the tissue distribution of coinjected dye (Fig. 5). Histological analysis of coronal sections through PVN revealed that injections were typically centered near the rostral-caudal midpoint of PVN. In most cases, dye extended rostrally and caudally to include the entire PVN. One rat in each of the ATZ, H₂O₂, and PEG-catalase groups showed a minor encroachment rostrally into the caudal preoptic area. Data were included because drug effects in each case were indistinguishable from those where injectate was confined to PVN.

Fig. 5.

Histological verification of PVN injection sites. A: an unstained histological section through mid-PVN shows the distribution of blue dye (arrow) that marks the site of l-glu microinjection. Note that this section was from a representative ATZ experiment. B: schematic drawings of rostral (top), middle (center), and caudal (bottom) sections through PVN show estimates of l-glu distribution based on the distribution of Chicago sky blue dye (100 nl). Distributions were nearly identical among rats receiving the catalase inhibitor ATZ, H₂O₂, or PBS vehicle. Therefore, their dye distributions are indicated together by shaded regions on the right side. Shaded regions on the left side indicate estimated tissue distribution of l-glu in the PEG-cat and PEG vehicle groups. All injections were centered on PVN, with dye encroaching slightly lateral and ventral to PVN. Note: dye distribution might not precisely reflect the tissue distribution of PEG-cat due to the fact that the molecular mass of dye (0.99 kDa) is considerably less than that of PEG-cat (255 kDa). Also note that, because of their large molecular masses and limited tissue diffusion, a larger volume (200 nl) of PEG-cat/PEG vehicle was injected into PVN compared with other test compounds (100 nl). Numbers to the right of each panel indicate stereotaxic distance from bregma.

DISCUSSION

Neurons of the hypothalamic PVN receive both intrinsic and extrinsic glutamatergic input (7, 25, 47), and exaggerated glutamatergic drive of PVN has been reported to support SNA under physiological and cardiovascular/metabolic disease conditions (18–20, 27, 29, 32, 39). ROS, including the membrane-permeable H₂O₂, can negatively and positively modulate glutamatergic neurotransmission (1–3, 48, 61), and here we provide evidence that endogenous H₂O₂ in PVN has a net inhibitory effect on renal sympathoexcitatory responses to stimulation with l-glu. Evidence is also provided that indicates that tonic inhibition by PVN H₂O₂ is submaximal due to ongoing enzymatic “scavenging” by catalase.

In this study, we injected l-glu unilaterally into PVN at a dose that acutely increased RSNA without changing MAP or HR. Our rationale was to minimize changes of synaptic input to PVN arising from arterial baroreceptors and possibly other visceral (sympathetic/vagal) afferents (15–17, 53). The lack of an increase of MAP likely reflects the anatomy of descending PVN presympathetic pathways, which mainly travel to the brain stem and spinal intermediolateral cell column on the ipsilateral side. Thus SNA to tissues on the side opposite of our PVN injection site likely increased far less than on the ipsilateral side where RSNA recordings were made. It is also possible that the transient increase of RSNA produced by l-glu injections, combined with the relatively slow response time of the vasculature, could have contributed to the lack of a pressor response. We also cannot rule out the possibility that SNA to nonrenal tissues may have decreased thereby offsetting the pressor effectiveness of increased RSNA.

Because we stimulated PVN with the endogenous transmitter l-glu, effects of manipulating tissue H₂O₂ levels could reflect modulation of ionotropic (25) and/or metabotropic glutamate receptors (mGluR) expressed in PVN (32, 46). Based on available evidence, activation of NMDA receptors in PVN robustly increases SNA, whereas non-NMDA (AMPA/kainite) receptors mostly mediate neuroendocrine responses (9, 25). Thus increases of RSNA by l-glu in the present study are more likely to reflect NMDA receptor activation. A role for mGluR, however, cannot be ruled out, since agonist activation of group I mGluR in PVN has also been reported to increase RSNA in anesthetized rats (32).

To determine effects of H₂O₂ on RSNA responses to PVN l-glu, we adopted the strategy of inhibiting catalase activity locally by microinjection of the inhibitor ATZ and allowing the concentration of endogenously generated H₂O₂ to rise. Because ATZ alone was without effect on recorded variables, we were able to directly and selectively evaluate modulation of l-glu-driven increases of RSNA. Lack of effect of ATZ alone is consistent with evidence that neither activity of presympathetic neurons, nor glutamatergic tonus of PVN normally contribute to ongoing sympathetic activity or maintenance of ABP or HR in anesthetized rats (15, 17, 20, 43). It is worth noting that our results are consistent with a previous study in which intracerebroventricular injections of ATZ and H₂O₂ also failed to change resting MAP and HR (28).

To interpret effects of ATZ, it is important to understand its mechanism of action. Catalase-mediated dismutation of H₂O₂ is normally divided into two steps, each consuming one molecule of H₂O₂. First, H₂O₂ oxidizes endogenous catalase, and then a second H₂O₂ regenerates catalase + 2 H₂O and O₂ (4, 54). ATZ reacts with the oxidized form of catalase and thus competes with the second step: regeneration of catalase (4, 24, 37). As a result, onset and efficacy of catalase inhibition by ATZ depends on the local concentration of H₂O₂. Because inhibition of catalase by ATZ was effective in the present study, it appears that catalase is constitutively active and that H₂O₂ is constitutively produced in PVN. Although basal PVN H₂O₂ concentration was not determined in our experiments and efficacy of ATZ to increase tissue H₂O₂ concentration was not quantified, we consider it likely that the effect of ATZ was mediated by catalase inhibition, since the dose used (25 nmol) has been previously reported to effectively inhibit brain catalase activity (4, 53).

To assess whether effects of ATZ were consistent with a rise of PVN H₂O₂, we compared its effects to those of exogenous H₂O₂. Like ATZ, H₂O₂ injection into PVN did not change resting RSNA, MAP, or HR, but did reduce the l-glu-evoked increase of RSNA. Thus effects of PVN ATZ appear consistent with those expected from an increase of PVN H₂O₂. It is noteworthy that PVN microinjection of ATZ was previously reported to not affect acute sympathoexcitatory responses to local microinjection of angiotensin II (ANG II) (51). Therefore, it seems that the inhibitory effect of ATZ is not likely attributable to nonspecific depression of PVN neuronal excitability. Taken together with the present observation that ATZ and H₂O₂ attenuate increases of RSNA elicited by l-glu, it seems that inhibitory modulation might be specific to neuronal excitatory actions of l-glu.

We next addressed the question of whether ongoing endogenous production of H₂O₂ is sufficient to tonically suppress the increase of RSNA evoked by PVN l-glu. This was accomplished by delivering a lipophilic form of catalase (PEG-catalase) into PVN to increase enzymatic scavenging of H₂O₂, thereby diminishing tonic effects of H₂O₂. PEG-catalase was delivered unilaterally to PVN and the dose (1.0 IU) was at least fivefold greater than the submaximal doses previously used in studies of PVN (51, 53). We determined that our dose of PEG-catalase alone had no effect on RSNA, MAP, or HR, suggesting that, if ongoing catalase activity in PVN was insufficient to fully scavenge endogenously produced H₂O₂, then any excess H₂O₂ was insufficient to tonically influence recorded variables. The latter could reflect the fact that PVN was targeted unilaterally, since bilateral PVN injection of a smaller dose of PEG-catalase has been reported to decrease ongoing RSNA in anesthetized rats (52).

Our observation that the increase of RSNA by PVN l-glu was enhanced by injection of PEG-catalase suggests that basal and/or l-glu-stimulated H₂O₂ in PVN exceeds the capacity of enzymatic scavenging by endogenous catalase. This conclusion is consistent with evidence from studies of dopamine neurons (41), but contrasts with recent evidence in PVN (52), where PEG-catalase was reported to have no effect on RSNA and pressor responses to l-glu. There appear to be two likely explanations for this discrepancy. First, our study tested effects of PEG-catalase on l-glu-evoked responses after only ∼1 min had elapsed, whereas 20 min separated injections of PEG-cat and l-glu in the earlier study. As mentioned above, our dose of PEG-catalase was also fivefold greater than that in the previous study by Xu et al. (52), and this could have also contributed to our detection of an enhanced RSNA response to PVN l-glu.

A caveat to the above interpretation is that experiments were performed in anesthetized rats, where delivery of oxygen to brain may be compromised. Under such conditions, the relative balance of H₂O₂ production (basal or l-glu stimulated) vs. scavenging might be shifted to favor oxidative “stress”. To mitigate this potential confound, rats were artificially ventilated with supplemental O₂. Nevertheless, additional experiments are needed in conscious rats to confirm whether H₂O₂ is a physiological regulator of PVN-mediated control of sympathetic activity. Additional studies are also needed to determine specific physiological or disease conditions in which inhibitory regulation by H₂O₂ of l-glu-mediated sympathetic activation might become enhanced or diminished.

Although sources of H₂O₂ in PVN remain to be determined, perhaps the most likely is dismutation of O₂^·− by SOD. In neurons, stimulation with l-glu can induce a short latency mitochondrial oxidative burst that generates O₂^·− and H₂O₂ (41). A similar action could occur in PVN, given its dense innervation by glutamatergic inputs (7, 31). Another possible source of H₂O₂ in PVN is the neuropeptide ANG II. In subfornical organ neurons, ANG II stimulates O₂^·− via NAD(P)H oxidase (58), which has been implicated in ANG II-induced neuronal excitation and development of arterial hypertension (59). In PVN, ANG II is also neuronal excitatory (11, 19, 30) and has been linked to production of H₂O₂ (51, 53). In contrast to the present study, however, H₂O₂ in PVN has been previously reported to enhance the renal sympathoexcitatory response to activation of the cardiac sympathetic afferent reflex (51, 53), rather than attenuate l-glu-evoked sympathetic activation.

In summary, endogenously generated H₂O₂ in PVN is sufficient under conditions of these experiments to suppress l-glu-evoked activation of renal sympathetic outflow. Tonic inhibitory effects of PVN H₂O₂ appear submaximal, since endogenously generated or exogenously delivered H₂O₂ can further attenuate sympathetic activation by l-glu. Regardless of whether H₂O₂ levels in PVN are stable or can be stimulated by glutamatergic input, ANG II, or other neuronal excitatory mediators, H₂O₂ in PVN has the capacity to buffer neuronal excitation by l-glu. This could be important for preventing hyperactivation of sympathoexcitatory PVN neurons, as well as their downstream synaptic targets. Given the potential for H₂O₂ to impact regulation of sympathetic outflow, additional studies are needed at the single-cell level to determine its mechanism of action.

GRANTS

This research was supported by Fundacao de Amparo a Pesquisa do Estado de Minas Gerais and Coordenacao de Aperfeicoamento de Pessoal de Nivel Superior to L. M. Cardoso, Conselho Naciona de Pesquisa to L. M. Cardoso and E. Colombari. Fundacao de Amparo a Pesquisa do Estado de Sao Paulo to E. Colombari, and National Heart, Lung, and Blood Institute Grants R01 HL 102310 and PO1 HL088052 to G. M. Toney.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: L.M.C., E.C., and G.M.T. conception and design of research; L.M.C. performed experiments; L.M.C. and G.M.T. analyzed data; L.M.C., E.C., and G.M.T. interpreted results of experiments; L.M.C. and G.M.T. prepared figures; L.M.C. drafted manuscript; L.M.C., E.C., and G.M.T. approved final version of manuscript; E.C. and G.M.T. edited and revised manuscript.