The proinflammatory cytokine tumor necrosis factor-α excites subfornical organ neurons

Considerable recent evidence has suggested important links between inflammation and the pathological mechanisms underlying hypertension. The present study describes cellular mechanisms through which acute and long-term exposure of tumor necrosis factor-α (TNF-α) influences the activity of subfornical organ neurons by modulating the voltage-gated transient Na⁺ current. This provides critical new information regarding the specific pathological mechanisms through which inflammation and TNF-α in particular may result in the development of hypertension.

Abstract

Tumor necrosis factor-α (TNF-α) is a proinflammatory cytokine implicated in cardiovascular and autonomic regulation via actions in the central nervous system. TNF-α^−/− mice do not develop angiotensin II (ANG II)-induced hypertension, and administration of TNF-α into the bloodstream of rats increases blood pressure and sympathetic tone. Recent studies have shown that lesion of the subfornical organ (SFO) attenuates the hypertensive and autonomic effects of TNF-α, while direct administration of TNF-α into the SFO increases blood pressure, suggesting the SFO to be a key site for the actions of TNF-α. Therefore, we used patch-clamp techniques to examine both acute and long-term effects of TNF-α on the excitability of Sprague-Dawley rat SFO neurons. It was observed that acute bath application of TNF-α depolarized SFO neurons and subsequently increased action potential firing rate. Furthermore, the magnitude of depolarization and the proportion of depolarized SFO neurons were concentration dependent. Interestingly, following 24-h incubation with TNF-α, the basal firing rate of the SFO neurons was increased and the rheobase was decreased, suggesting that TNF-α elevates SFO neuron excitability. This effect was likely mediated by the transient sodium current, as TNF-α increased the magnitude of the current and lowered its threshold of activation. In contrast, TNF-α did not appear to modulate either the delayed rectifier potassium current or the transient potassium current. These data suggest that acute and long-term TNF-α exposure elevates SFO neuron activity, providing a basis for TNF-α hypertensive and sympathetic effects.

NEW & NOTEWORTHY Considerable recent evidence has suggested important links between inflammation and the pathological mechanisms underlying hypertension. The present study describes cellular mechanisms through which acute and long-term exposure of tumor necrosis factor-α (TNF-α) influences the activity of subfornical organ neurons by modulating the voltage-gated transient Na⁺ current. This provides critical new information regarding the specific pathological mechanisms through which inflammation and TNF-α in particular may result in the development of hypertension.

hypertension is the leading risk factor for the development of cardiovascular disease (Lim et al. 2012). While the consequences of hypertension are well understood, its etiology remains elusive. One line of thought suggests that autonomic nervous system dysfunction may contribute to the development of hypertension by increasing sympathetic output and inducing the release of vasoconstricting agents such as catecholamines and angiotensin II (ANG II; Seravalle et al. 2014). Interestingly, elevated sympathetic nerve activity has been observed in patients with borderline hypertension, suggesting that autonomic dysfunction may play an important role in the development of hypertension (Anderson et al. 1989; Julius et al. 1991).

Tumor necrosis factor-α (TNF-α), a proinflammatory cytokine, is elevated in the plasma of patients with hypertension and cardiovascular disease (Aukrust et al. 1998; Glezeva and Baugh 2014, Tedgui and Mallat 2006). Recent evidence indicates that TNF-α acts within the central nervous system (CNS) to increase blood pressure and activate key autonomic centers, such as the paraventricular nucleus (PVN) of the hypothalamus (Zhang et al. 2003). Strikingly, TNF-α^−/− mice do not develop hypertension or cardiac hypertrophy (Sriramula et al. 2008a) in an ANG II-salt model of hypertension. These data implicate the involvement of TNF-α in blood pressure regulation and sympathetic output. However, CNS sites at which TNF-α may exert its sympathetic and pressor effects have only been recently explored.

The subfornical organ (SFO) is a circumventricular forebrain structure that lacks a complete blood-brain barrier (BBB), making it an ideal site for sensing circulating signals in the periphery that do not diffuse across the normal BBB (Ganong 2000; Smith and Ferguson 2010). It has been well established that the SFO serves as a fundamental site for fluid homeostasis and cardiovascular regulation due to its ability to mediate ANG II dipsogenic (Moreau et al. 2012; Simpson and Routtenberg 1973; Tanaka et al. 2003) and pressor effects (Mangiapane and Simpson 1980). Strikingly, the SFO is largely responsible for the induction of hypertension by ANG II-salt loading (Osborn et al. 2012), which signifies that it may be a site at which pathological mechanisms of hypertension may converge. Electrical stimulation of the SFO increases blood pressure (Mangiapane and Brody 1983) and induces vasopressin release by the PVN (Ferguson and Kasting 1986). Moreover, efferent SFO projections innervate PVN neurons that project to the dorsal medulla, a critical site for sympathetic output (Ferguson et al. 1984). Neurons within the SFO express TNF-α receptor 1 (TNF-R1; Hindmarch and Ferguson 2016), and intriguingly, it is colocalized with the type 1 ANG II receptor (AT₁R; Wei et al. 2015). It has been determined that injection of TNF-α into the carotid artery of rats increases PVN action potential firing rate, sympathetic nerve activity, and arterial pressure (Zhang et al. 2003), and the effects on sympathetic nerve activity and arterial pressure are attenuated by lesioning the SFO (Wei et al. 2013). This suggests that TNF-α may have direct actions on the SFO. In agreement with this hypothesis, it was found that direct microinjection of TNF-α into the SFO dramatically elevates blood pressure and expression of AT₁R in both the SFO and PVN (Wei et al. 2015). Together, these findings suggest that the SFO plays an important role in the pressor and sympathetic effects of TNF-α. However, the underlying mechanisms by which TNF-α influences the SFO to mediate these effects are currently unknown.

The goal of the present study was to investigate the effects of TNF-α on the activity of SFO neurons. To test this, we performed whole cell and perforated patch-clamp experiments on dissociated SFO neurons in the presence of TNF-α and observed changes in activity and membrane excitability. Collectively, our data suggest that acute and long-term application of TNF-α excites SFO neurons, and the long-term effects are mediated by the voltage-gated transient sodium current (I_Na).

MATERIALS AND METHODS

Animals

Male Sprague-Dawley rats (Charles River, Quebec, Canada) that were 28–35 days old and weighed between 100 and 150 g were used for all experiments. Rats were housed at 22°C under a 12-h light-dark cycle and were provided food and water ad libitum. All animal procedures were approved and in compliance with the Queen’s University Animal Care Committee, as well as with the Canadian Council on Animal Care guidelines.

Chemicals and Drugs

All chemicals and drugs were obtained from Sigma-Aldrich (Oakville, ON) unless otherwise stated. Tetrodotoxin was obtained from Alomone Laboratories (Jerusalem, Israel) and stored at −20°C. TNF-α was obtained from Phoenix Pharmaceuticals (Burlingame, CA) and stored at −20°C. All drugs were prepared in artificial cerebrospinal fluid (aCSF) or intracellular solution the day of experimentation.

SFO Neuron Preparation

Dissociated SFO neurons were prepared as previously described (Ferguson et al. 1997; Kuksis and Ferguson 2015). Rats were euthanized by decapitation and their brains were dissected out and placed in ice-cold, carbogenated (95% O₂-5% CO₂) aCSF containing the following (in mM): 124 NaCl, 20 NaHCO₃, 10 glucose, 2.27 CaCl_2, 2.5 KCl, 1.3 MgSO₄, and 1.24 KH₂PO₄. The SFOs from three euthanized rats were dissociated and cultured together each week for experimentation. The SFOs were then microdissected and placed in a drop of Hibernate media (ThermoFisher Scientific, Burlington, ON, Canada) containing B27 supplement (ThermoFisher Scientific). After dissection, the SFOs were incubated in 5 ml of Hibernate media containing 10 mg papain (Worthington Biochemical, Lakewood, NJ) for 30 min at 30°C in a warm water bath to allow for digestion of the extracellular matrix proteins. The SFOs were then rinsed and triturated three times each with Hibernate containing B27 to dissociate the SFOs into individual cells. The resulting cell suspension was centrifuged at 200 g for 8 min. The supernatant was removed and the pellet was resuspended in 30 μl Neurobasal-A media (ThermoFisher Scientific) containing B27, 100 U/ml penicillin-streptomycin (ThermoFisher Scientific), 0.4 mM l-glutamine (ThermoFisher Scientific), and 5 mM glucose. Ten-microliter drops were plated individually on 35 × 10 mm polystyrene dishes (Corning Life Sciences, Corning, NY) and were incubated at 37°C in 5% CO₂ for 1.5 h to allow the cells to adhere to the dish. Following the incubation period, 2 ml of Neurobasal-A media were added to the dishes and the cells were incubated for at least 24 h before experiments were performed. Twenty-four hours before experiments on neurons incubated in TNF-α, the Neurobasal-A media were replaced with a Neurobasal-A media containing 10 ng/ml (575 pM) TNF-α, a concentration that was chosen based on a previous study showing the effects of TNF-α on mouse cortical neurons (Chen et al. 2015).

Electrophysiology

Whole cell and perforated patch-clamp recordings were performed on SFO neurons using the Multiclamp 700B patch-clamp amplifier (Molecular Devices, Sunnyvale, CA) and were filtered at 2 kHz and digitized to 5 kHz. Spike2 (v7) and Signal (v6) software programs (Cambridge Electronics Design, Cambridge, UK) were used to record the data and analyze it post hoc. Micropipettes of 3–5 MΩ were prepared using borosilicate glass (World Precision Instruments, Sarasota, FL) using a P47 Flaming-Brown Micropipette Puller (Sutter Instrument, Novato, CA). The dishes containing SFO neurons were perfused at a rate of 2 ml/min with 37°C aCSF that contained the following (in mM): 140 NaCl, 10 HEPES, 5 KCl, 5 glucose, 5 mannitol, 2 CaCl₂, and 1 MgCl₂ (280–300 mosM) and had a pH adjusted to 7.3 using NaOH. Perfusion was performed using a gravity perfusion system, and the flow of aCSF was established with a vacuum pump. A MP-225 micromanipulator (Sutter Instrument) was used to guide the micropipette containing the electrode and intracellular solution toward a SFO cell membrane. Once a membrane was reached, negative pressure was applied to allow the formation of a >1-GΩ seal. For whole cell patch-clamp experiments, further negative pressure was applied until whole cell access was achieved. For perforated patch-clamp experiments, time was given for the Amphotericin B (400 μg/ml, 1% DMSO) within the micropipette intracellular solution to form pores in the membrane until adequate whole cell access was achieved (Linley 2013). It was determined whether the cell was a neuron based on the presence of a fast-activating voltage-gated inward conductance during a depolarizing voltage-step protocol. For voltage-clamp experiments, only recordings with a series resistance (R_s) of 5–20 MΩ were kept for analysis.

Whole cell and perforated current-clamp experiments.

For whole cell and perforated current-clamp experiments, the aCSF described above was used unless otherwise specified. An intracellular solution containing the following (in mM): 125 K-gluconate, 10 KCl, 10 HEPES, 2 MgCl₂, 2 Na₂ATP, 1.1 EGTA, and 0.3 CaCl₂ (270–290 mosM) was used for current-clamp experiments. This intracellular solution pH was adjusted to 7.3 using KOH and had a calculated liquid junction potential of −15.0 mV. To measure acute changes in membrane potential, 4 ml TNF-α (5 fM to 5 nM) were bath perfused into the dish. Before bath application of TNF-α, a >100-s baseline was established and neurons were only analyzed whether action potentials were >60 mV. Each dish was tested with only a single bath application of TNF-α. To measure firing rate, the neurons were held at −65 mV with a holding current and the firing rate was obtained from a 100-s bin. The rheobase for each neuron was obtained by resting them at a membrane potential of −80 mV with a holding current to prevent spontaneous activity, and current pulses (500 ms) of increasing magnitude were applied until one action potential was observed.

Isolation and measurement of the transient sodium current.

To isolate the I_Na during whole cell voltage-clamp experiments, modified aCSF and intracellular solution were used to remove the contribution of K⁺ and Ca²⁺ currents. The modified aCSF contained the following (in mM): 110 NaCl, 20 TEA-Cl, 10 HEPES, 5 glucose, 5 CsCl, 1 MgCl₂, and 0.3 CdCl₂ (280–300 mosM), and its pH was adjusted to 7.3 using NaOH. In addition, the modified intracellular solution contained the following (in mM): 125 CsMeSO₄, 10 CsCl, 10 HEPES, 2 MgCl₂, 2 Na₂ATP, 1.1 EGTA, and 0.3 CaCl₂ (270–290 mosM), and its pH was adjusted to 7.3 using CsOH. This intracellular solution had a liquid junction potential of −18.1 mV. To measure the activation kinetics and the peak magnitude of the I_Na, a voltage-step protocol was used. The neurons were held at −75 mV, and then a prestep was performed to −90 mV for 100 ms, followed by 250-ms voltage steps from −95 to −35 mV in 5-mV increments. Peak inward currents for each voltage step were measured and conductance values (G_Na) were calculated using the equation below:

$$G_{\text{Na}}\text{=}\frac{I_{\text{Na}}}{\text{(}{V}_{\text{m}}-E_{\text{Na}}\text{)}}$$

where I_Na is the peak inward current, E_Na is the equilibrium potential for I_Na (+107.1 mV), and V_m is the voltage step that induced I_Na. Conductance values were normalized to the maximum conductance of that neuron, and the normalized conductance values were then plotted against the pertaining voltage step. To measure the I_Na inactivation kinetics, the neurons were clamped to −75 mV, followed by a series of presteps from −125 to −35 mV for 500 ms in 10-mV increments and succeeded by a voltage step to −25 mV for 100 ms. Just as was done for the activation kinetics of the I_Na, the peak inward currents corresponding to the −25-mV voltage step and its preceding prestep were measured and conductance values were calculated using the formula above. These conductance values were normalized to the maximum conductance of that neuron and were plotted against its corresponding voltage step.

Isolation and measurement of the delayed rectifier potassium current.

SFO neurons have two primary potassium currents that contribute to the majority of the outward current: the I_K and I_A (Anderson et al. 2001). These currents can be separated from one another using modified intracellular and extracellular solutions in combination with manipulation of cell membrane potential. For whole cell voltage-clamp experiments to isolate the I_K, modified aCSF and intracellular solution were used to remove the contribution of Na⁺ and Ca²⁺ currents. The modified aCSF contained the following (in mM): 140 NaCl, 5 KCl, 1 MgCl₂, 2 CaCl₂, 10 HEPES, 5 glucose, 5 mannitol, 0.3 CdCl₂, and 0.001 tetrodotoxin (280–300 mosM) and was adjusted to a pH of 7.3 using NaOH. The intracellular solution contained the following (in mM): 125 K-gluconate, 10 KCl, 10 HEPES, 5.5 EGTA, 2 MgCl₂, 2 Na₂ATP, and 0.3 CaCl₂ (270–290 mosM) and was adjusted to a pH of 7.3 using KOH. To measure the peak magnitude of the I_K, a voltage-step protocol was used that clamped the neurons at −75 mV, followed by a prestep to −40 mV for 100 ms, and then voltage steps from −80 to 20 mV for 250 ms in 10-mV increments. The prestep to −40 mV inactivates the I_A component of the total outward current. The peak outward magnitudes in the last 50 ms of the voltage step were taken and plotted against its corresponding voltage step.

Isolation and measurement of the transient potassium current.

For whole cell voltage-clamp experiments to isolate the I_A, the modified aCSF and intracellular solution that were used for the isolation of the I_K above were used here. A similar voltage-step protocol was used, but instead a prestep to −90 mV was performed to prevent the inactivation of the I_A. This voltage-step protocol was performed on the same neuron and in direct succession to the voltage-step protocol that was performed to isolate the I_K. The peak outward current magnitudes within the first 50 ms of the voltage steps were subtracted from the peak outward current magnitudes observed during the I_K voltage-step protocol, which generated a set of values pertaining to the I_A component of the total outward current. The I_A at each voltage step was normalized to the maximum I_A of that neuron and then plotted against its corresponding voltage step.

Statistical Analysis

All data are reported as means ± SE. Changes in membrane potential following acute bath application of TNF-α were deemed significant if the magnitude of the change was at least two times the SD of the resting membrane potential before application. All comparisons of the mean values between two groups were analyzed using unpaired Student’s t-tests. A Boltzmann function was used to calculate the V₅₀ and slope of the I_Na activation and inactivation profiles, as well as the I_A activation profile. The liquid junction potential was corrected for in all the data analysis and figures. All data analysis was performed using GraphPad Prism 6.07 (La Jolla, CA). The level of significance for all statistical tests was set at P < 0.05.

RESULTS

Acute Bath Application of TNF-α Depolarizes SFO Neurons

Based on previous reports that microinjection of TNF-α into the SFO elevates blood pressure and PVN firing rate (Wei et al. 2015), we hypothesized that TNF-α increases the excitability of SFO neurons to cause these effects. To determine the effects of TNF-α on the excitability SFO neurons, we conducted perforated current-clamp experiments on 62 dissociated SFO neurons obtained from 39 animals and observed changes in membrane potential and firing rate following bath application of TNF-α. We found that bath application of 5 nM TNF-α depolarized 50% (n = 5/10) of SFO neurons with a mean magnitude of depolarization of 6.4 ± 0.6 mV (n = 5; Fig. 1, A and B). As illustrated in Fig. 1, A and B, additional groups of SFO neurons were tested with TNF-α at concentrations between 5 fM and 5 nM, and collectively a total of 35% depolarized (n = 22/62). The depolarizations were typically accompanied by an increase in firing rate (Fig. 1C) and lasted around 10–30 min. It should be noted that some neurons, such as in Fig. 1A, bottom, exhibited a minimal or nonspontaneous firing rate before and during bath application of TNF-α. There was no significant difference in the resting membrane potential between nonresponsive and depolarized SFO neurons (nonresponsive = −56.5 ± 1.1 mV, n = 40; depolarized = −56.9 ± 1.5 mV, n = 22; P = 0.842), suggesting that TNF-α effects, or lack thereof, were not due to differences in the resting membrane potential. Interestingly, both the proportion of depolarized SFO neurons and the magnitude of depolarization induced by TNF-α were concentration dependent and had nearly identical EC₅₀ values (EC₅₀ = 1.0 ± 0.1 pM, and EC₅₀ = 0.9 ± 0.4 pM, respectively; Fig. 1B). These data suggest that TNF-α is able to acutely influence the activity of SFO neurons in a concentration-dependent manner.

Fig. 1.

Acute bath application of TNF-α depolarizes subfornical organ (SFO) neurons. A: representative current-clamp recordings from 2 dissociated SFO neurons highlighting that bath application of 5 nM (top) and 500 pM (bottom) TNF-α depolarizes SFO neurons. The blue bar denotes the time and duration of bath application of TNF-α, and the dashed red line denotes the baseline membrane potential before TNF-α bath application. B: proportion of depolarizing SFO neurons (black) and the magnitude of depolarization (blue) at each concentration are shown (n = 62). The data sets were superimposed with a sigmoid concentration-response function, and the EC₅₀ values are highlighted by the dashed vertical lines and as numerical values. C: action potential firing rate before and during bath application of 5 nM (blue), 500 pM (light blue), and 50 pM (black) TNF-α. The firing rates during bath application of TNF-α relative to the firing rates before for each concentration were significantly different (5 nM: Δfiring rate = 0.9 ± 0.3 Hz, *P = 0.04; 500 pM: Δfiring rate = 1.6 ± 0.7 Hz, *P = 0.05; 50 pM: Δfiring rate = 0.8 ± 0.2 Hz, *P = 0.02).

TNF-α Incubation Increases SFO Neuron Excitability

To determine the long-term effects of TNF-α on SFO neuron excitability, we incubated SFO neurons in 10 ng/ml (575 pM) TNF-α for 24 h and then performed perforated current-clamp experiments to examine their properties. We determined the magnitude of the minimum current required to induce a single action potential (rheobase) in neurons incubated in either TNF-α or control media (control; Fig. 2A). To do this, the SFO neuron membrane potential was held at −80 mV to allow for the measurement of membrane excitability without the presence of spontaneous activity. It was observed that neurons incubated in TNF-α exhibited a rheobase of smaller magnitude relative to control (control = 11.9 ± 1.7 pA, n = 8; TNF-α = 5.6 ± 0.5 pA, n = 10; P = 0.007; Fig. 2, A and B). Both the resting membrane potential and the input resistance were not significantly different between the control group and the TNF-α-treated group (resting membrane potential: control = −79.7 ± 0.4 mV, n = 8; TNF-α = −79.3 ± 0.4 mV, n = 10; P = 0.541; input resistance: control = 1.6 ± 0.2 GΩ, n = 8; TNF-α = 1.8 ± 0.3 GΩ, n = 10; P = 0.684). These data suggest that TNF-α increases SFO neuron membrane excitability and this effect was due to changes in the intrinsic properties of these neurons.

Fig. 2.

TNF-α incubation increases SFO neuron excitability. A: representative current-clamp traces of a control neuron (−80 mV baseline; top) and a neuron incubated in 10 ng/ml (575 pM) TNF-α (−80 mV baseline; middle) during current pulses (bottom) of 6 pA (left) and 12 pA (right). B: rheobase of control neurons (white) compared with neurons incubated in TNF-α (blue), highlighting that neurons incubated in TNF-α have a significantly smaller rheobase (**P = 0.007). C: representative current-clamp traces of the firing rate of control neurons (−65 mV; top) compared with neurons incubated in TNF-α (−65 mV; bottom) following no stimulus. D: basal firing rate of control neurons (white) and of neurons incubated in TNF-α (blue), highlighting that neurons incubated in TNF-α have a significantly higher firing rate (**P = 0.019).

Based on the observation that TNF-α incubation lowers the rheobase magnitude, we then hypothesized that incubation in TNF-α would increase the firing rate of SFO neurons. To test this we measured the firing rate of neurons incubated in TNF-α and compared with control. It was found that 70% (n = 7/10) of neurons in each group exhibited spontaneous action potentials (Fig. 2C). Of the neurons that were spontaneously active, the neurons incubated in TNF-α had a significantly elevated firing rate relative to that of control (control = 0.5 ± 0.2 Hz, n = 7; TNF-α = 3.0 ± 0.8 Hz, n = 7; P = 0.019; Fig. 2, C and D). The mean resting membrane potentials between the groups were not significantly different (control = −64.8 ± 0.5 mV, n = 7; TNF-α = −65.5 ± 0.8 mV, n = 7; P = 0.428), indicating that the difference in firing rate was not due to a difference in the membrane potential. Overall, these data suggest that TNF-α incubation increases SFO neuron membrane excitability and basal activity.

TNF-α Incubation Enhances the Transient Sodium Current

Based on the effects of TNF-α incubation on rheobase and firing rate, we hypothesized that TNF-α was potentiating the I_Na. To investigate this, we incubated neurons in 575 pM TNF-α and then performed whole cell voltage-clamp experiments on these neurons using a voltage-step protocol (see materials and methods). We explored the effects of TNF-α on the peak I_Na magnitude and found that neurons incubated in TNF-α had a significantly larger peak I_Na magnitude relative to control (control = 1.7 ± 0.1 nA, n = 15; TNF-α = 2.7 ± 0.2 nA, n = 12; P < 0.001; Fig. 3, A and B). Intriguingly, we also observed that neurons incubated in TNF-α exhibited a hyperpolarized V₅₀ compared with control (control = −45.2 ± 0.4 mV, n = 15; TNF-α = −49.4 ± 0.3 mV, n = 12; P < 0.001; Fig. 3C), but TNF-α had no effect on the slope of the activation curve (control = 4.0 ± 0.4, n = 15, TNF-α = 3.3 ± 0.2, n = 12; P = 0.147). Thus TNF-α lowered the membrane potential at which 50% of the I_Na activates, which may explain why TNF-α incubation elevated the firing rate and decreased the rheobase. Next, we investigated the effects of TNF-α on the inactivation kinetics of I_Na and found that TNF-α incubation did not influence I_Na inactivation (V₅₀: control = −73.2 ± 1.0 mV, n = 11; TNF-α = −73.4 ± 0.8 mV, n = 16; P = 0.889; slope: control = −10 ± 1.0, n = 11; TNF-α = −8.8 ± 0.7, n = 16; P = 0.323; Fig. 3D). Taken together, these data suggest that TNF-α incubation increases the excitability of SFO neurons by enhancing the I_Na.

Fig. 3.

TNF-α enhances the transient sodium current. A: representative overlaid voltage-clamp traces of the sodium current (I_Na) during each voltage step in control neurons (left) compared with neurons incubated in TNF-α (right), both of which were performed in the presence of extracellular and intracellular solutions that blocked K⁺ and Ca²⁺ currents. B: peak I_Na magnitude of control neurons (white) compared with neurons incubated in TNF-α (blue). C: normalized conductance at each voltage step of the activation protocol was plotted for both control neurons (black) and neurons incubated in TNF-α (blue). The data sets were fitted with a Boltzmann sigmoidal function and the representative curves were overlaid. The voltage (V₅₀) for the data sets is highlighted by the dashed vertical lines and the numerical values (****P < 0.001). Inset: overlapping window currents for the activation and inactivation curves of TNF-α (light blue and dark grey) and control (light grey and dark grey). D: normalized conductance at each voltage step of the inactivation protocol was plotted for both control neurons (black) and neurons incubated in TNF-α (blue). The data sets were fitted with a Boltzmann sigmoidal function and the representative curves were overlaid. The V₅₀ for the data sets is highlighted by the dashed vertical lines and the numerical values. G_Na, conductance values.

TNF-α Has No Effect on the Delayed Rectifier Potassium Current or the Transient Potassium Current

We next investigated whether voltage-gated potassium currents in SFO neurons were also modulated by incubation in TNF-α. We observed no significant difference between peak magnitude of the I_K in neurons incubated in 575 pM TNF-α compared with control (control = 1.1 ± 0.1 nA, n = 11; TNF-α = 1.2 ± 0.1 nA, n = 11; P = 0.713; Fig. 4, A–C). Although the V₅₀ or the slope could not be reported (as the I_K magnitude did not reach a plateau at the larger voltage steps), we compared the current magnitudes at each voltage step between the two groups and found no significant differences (Fig. 4C).

Fig. 4.

TNF-α has no effect on the delayed rectifier potassium current. A: representative voltage-clamp traces of the I_K in control neurons (left) and neurons incubated in TNF-α (right) following a voltage step to 20 mV. These recordings were performed in the presence of extracellular and intracellular solutions that blocked Na⁺ and Ca²⁺ currents. The black arrows highlight the very small tail currents observed when voltage is stepped back to −75 mV (close to reversal potential of −82 mV). B: peak I_K magnitude at 20 mV for control neurons (white) and neurons incubated in TNF-α (blue). C: magnitude of the I_K at each voltage step for control neurons (black) and neurons incubated in TNF-α (blue). Individual data points within each data set were connected with a straight line between each data point.

Similarly, we tested whether TNF-α incubation had an effect on the I_A (Fig. 5A). It was determined that TNF-α incubation had no significant effect on the peak magnitude of the I_A (control = 0.6 ± 0.1 nA, n = 11; TNF-α = 0.6 ± 0.1 nA, n = 8; P = 0.893; Fig. 5B). Furthermore, TNF-α incubation had no significant effect on the activation kinetics of the I_A (V₅₀: control = −8.4 ± 3.6 mV, n = 11; TNF-α = −11.4 ± 2.0 mV, n = 7; P = 0.546; slope: control = 14.6 ± 2.8, n = 11; TNF-α = 14.7 ± 1.7, n = 7; P = 0.977; Fig. 5C). These data suggest that TNF-α incubation has no effect on the I_K or the I_A, and therefore, these currents do not contribute to TNF-α effects on SFO neuron excitability.

Fig. 5.

TNF-α has no effect on the transient potassium current. A: representative voltage-clamp traces of the I_A in control neurons (left) and neurons incubated in TNF-α (right) following a voltage step to 20 mV. These recordings were performed in the presence of extracellular and intracellular solutions that blocked Na⁺ and Ca²⁺ currents. B: peak I_A magnitude at 20 mV for control neurons (white) and neurons incubated in TNF-α (blue). C: normalized current values at each voltage step of the activation protocol for the I_A was plotted for both control neurons (black) and neurons incubated in TNF-α (blue). The data sets were fitted with a Boltzmann sigmoidal function and the representative curves were overlaid. The V₅₀ for the data sets is highlighted by the dashed vertical lines and the numerical values.

DISCUSSION

In the present study, we investigated the effects of acute and long-term TNF-α exposure on the excitability of SFO neurons and the potential mechanisms underlying these long-term effects. We found that TNF-α depolarizes SFO neurons in a concentration-dependent manner and these depolarizations were accompanied by an increase in firing rate. Following an incubation period with TNF-α, we identified an elevated basal firing rate and a decrease in rheobase in SFO neurons, suggesting an increase in overall membrane excitability. Furthermore, TNF-α incubation increased the magnitude of the I_Na and shifted the activation curve of the I_Na to a more hyperpolarized state. However, TNF-α incubation had no effect on magnitude and activation kinetics of the I_K or the I_A. The concentrations of TNF-α used in the present in vitro study are higher than normal physiological ranges and at best reflect high pathophysiological concentrations previously reported in humans and animals (Drímal et al. 2008; Maury and Teppo 1989). Therefore, caution should be taken in the interpretation and application of these findings. Future studies will be necessary to determine whether these observations translate when physiological concentrations of TNF-α are induced in intact animal models of inflammation. Despite this limitation, our data show that acute and long-term application of TNF-α excites SFO neurons and the long-term effects are mediated by modulation of the I_Na but not the I_K or the I_A.

The Role of TNF-α in Central Cardiovascular and Sympathetic Regulation

The current study provides evidence for excitatory actions of TNF-α in the SFO, providing a basis for its central pressor and sympathetic effects. The SFO acts in a circuit with critical autonomic centers of the brain, such as the PVN, to control blood pressure and sympathetic output (Smith and Ferguson 2010). The consequences of activating the SFO include elevating blood pressure (Mangiapane and Brody, 1983), inducing vasopressin release (Ferguson and Kasting 1986), and increasing renal sympathetic nerve activity (Llewellyn et al. 2012). The SFO is known to project to the PVN, and efferent SFO fibers innervate distinct PVN neuron subtypes that regulate neuroendocrine and autonomic output (Kawano and Masuko 2010). Furthermore, ANG II, which has excitatory effects on SFO (Anderson et al. 2001; Cancelliere and Ferguson 2017) and PVN neurons (Li and Ferguson 1993a), has been shown to be released as a neurotransmitter from SFO efferents onto PVN neurons (Bains and Ferguson 1994; Li and Ferguson 1993b). Therefore, activation of the SFO may lead to synaptic release of ANG II (and potentially other neurotransmitters) within the PVN, and this may lead to an increase in blood pressure and sympathetic output. Direct administration of TNF-α into the SFO elevates the PVN firing rate, sympathetic nerve activity, and blood pressure (Wei et al. 2013). In combination with our data, these results implicate TNF-α as a modulator of this circuit and thus a modulator of neuroendocrine and autonomic activity.

Despite this, the time courses at which these effects take place remain to be elucidated. In the present study, the acute effects of TNF-α on SFO neurons typically lasted 10–30 min, whereas the increase in blood pressure caused by direct administration of TNF-α into the SFO persisted beyond 4 h (Wei et al. 2015). A possible explanation for these seemingly inconsistent observations may be that acute activation of the SFO by TNF-α may potentiate the activity of the downstream pathways that regulate blood pressure and sympathetic activity in the long term to allow for a constant and prolonged pressor response. Electrical stimulation of the SFO causes immediate increases in blood pressure that terminate immediately after cessation of stimulation (Mangiapane and Brody 1983), while administration of TNF-α into the SFO increases expression of TNF-R1 as well as AT₁R in the SFO and PVN (Wei et al. 2015). Overall, these data suggest that TNF-α is not just able to increase SFO neuron activity acutely but also potentiates the circuit to allow for a prolonged pressor response. It should also be considered that the routes of administration (i.e., bath application in the present study and microinjection in Wei et al. 2015) and the concentration of TNF-α (5 fM to 5 nM for bath application and 7 µM for microinjection) were different between the studies, which may potentially explain the discrepancy in the latency of sympathetic activation seen following microinjection of TNF-α. Lastly, it is known that TNF-α signaling induces reactive oxygen species generation (Barth et al. 2009), a process that is important for the pressor effects of ANG II (Zimmerman et al. 2004) and also has effects on other cell types, such as microglia and astroglia (Olmos and Lladó 2014). Perhaps these mechanisms allow for further neuronal activation in vivo on a longer time scale, which may contribute to the increased sympathetic activity seen following TNF-α administration into the SFO. Therefore, future studies should explore the contributions of these other potential actions of TNF-α within the SFO to sympathetic activation and blood pressure regulation. It remains unknown how the pressor and sympathetic effects of TNF-α translate into a pathological state such as hypertension. Future studies should investigate the potential pathological manifestations of TNF-α effects within the SFO.

A further consideration to our studies relates to the fact that all of our recordings were obtained from SFO neurons derived from male rats. Similarly, all of the work to date examining TNF-α action in SFO of intact animals has been carried out in males. There is now an extensive literature describing sex differences in many animal models of hypertension (Hay et al. 2014), including ANG II-induced hypertension (Xue et al. 2005, 2013), where estrogen receptors (Xue et al. 2015) and endoplasmic reticulum stress (Dai et al. 2016; Young et al. 2012) in the SFO have been implicated. While it is beyond the scope of the present study, in the future it will be intriguing to determine whether there are sex-related differences in the responsiveness of SFO neurons to TNF-α.

TNF-α and ANG II Cross Talk

ANG II is a peptide that plays a critical role in cardiovascular and autonomic regulation and is implicated in cardiovascular disease (Kishi 2012; Lastra and Sowers 2013). Centrally, ANG II is known to act at the SFO to increase blood pressure (Mangiapane and Simpson 1980). There appears to be extensive cross talk between TNF-α and ANG II in regards to cardiovascular regulation, primarily through an interaction within the PVN (Sriramula et al. 2013; Kang et al. 2008) and the SFO (Wei et al. 2015). It has been shown that TNF-α^−/− mice do not develop ANG II-induced hypertension or cardiac hypertrophy (Sriramula et al. 2008a). Intriguingly, ANG II-induced hypertension is mediated in large part by the SFO, as lesioning of this structure attenuates the pressor effect of ANG II (Osborn et al. 2012). Following administration of TNF-α into the SFO, AT₁R and other components of the renin-angiotensin system, such as angiotensin converting enzyme, are upregulated in the SFO and PVN (Wei et al. 2015). Furthermore, TNF-R1 and AT₁R are colocalized in the same SFO neurons (Wei et al. 2015). In the present study, TNF-α incubation of SFO neurons for 24 h (which attempts to mimic elevated circulating TNF-α during an immune challenge and in pathophysiological states like hypertension and heart failure in which the immune system is chronically activated) increases their membrane excitability and basal activity. In conjunction, these data suggest that 1) TNF-α is a vital component to central pressor effects of ANG II, and 2) increased TNF-α activity within the SFO modulates ANG II activity. One could therefore speculate that a pathological consequence of long-term increases in the circulating proinflammatory cytokine TNF-α may be priming the SFO to be more responsive to ANG II, thus allowing ANG II to induce a potentiated pressor response and subsequent hypertension. It has been previously shown that proinflammatory cytokines potentiate the pressor response to central ANG II (Ufnal et al. 2006; Zera et al. 2008; Sriramula et al. 2008b), but the mechanisms underlying these effects remain unknown. Future studies should examine whether previous TNF-α exposure potentiates the responsiveness of SFO neurons to ANG II and the signaling pathways by which this may occur.

TNF-α Effects on Ion Currents

Only recently have the central effects of TNF-α been explored. TNF-α is known to play a role in many neuronal processes including synaptic plasticity (Liu et al. 2017), neurodegeneration (Montgomery and Bowers 2012), and modulation of ion currents (Vezzani and Viviani 2015). In the present study, we identified the I_Na as a potential current underlying the long-term excitatory effects of TNF-α. A previous study in mouse cortical neurons showed that following TNF-α incubation nearly identical effects were seen on rheobase, firing rate, and I_Na kinetics when compared with the present study (Chen et al. 2015). On the other hand, TNF-α appeared to have no effect on voltage-gated K⁺ currents in other neuron types (Czeschik et al. 2008). These data are consistent with the data presented in the current study. What remains unknown is how TNF-α acutely depolarizes SFO neurons. Intriguingly, interleukin-1β (IL-1β), another proinflammatory cytokine, has similar depolarizing effects on SFO neurons that were shown to be mediated by a nonselective cationic conductance (Desson and Ferguson 2003). TNF-R1 and the IL-1β receptor share similar downstream signaling pathways in neurons (see Vezzani and Viviani 2015 for review), and thus, it is possible that TNF-α may also influence a nonselective cationic conductance. Future studies examining the mechanism(s) behind the depolarization induced by TNF-α may bring insight into how TNF-α elevates blood pressure and sympathetic activity.

In conclusion, we demonstrated that TNF-α increases the excitability of SFO neurons acutely and in the long term and that the long-term effects are mediated by the I_Na, thus implicating the SFO as a putative site for TNF-α to exert its central effects. In conjunction with previous findings, the actions of TNF-α at the SFO may promote activation of downstream autonomic and cardiovascular centers of the brain, such as the PVN, to elevate blood pressure and sympathetic tone. Our data, in combination with other studies, set a basis for future research to identify whether and how the effects of TNF-α in this circuit can pathologically manifest into hypertension.

GRANTS

This research was supported by the Canadian Institutes of Health Research.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

N.J.S. and A.V.F. conceived and designed research; N.J.S. performed experiments; N.J.S. and A.V.F. analyzed data; N.J.S. and A.V.F. interpreted results of experiments; N.J.S. prepared figures; N.J.S. and A.V.F. drafted manuscript; N.J.S. and A.V.F. edited and revised manuscript; N.J.S. and A.V.F. approved final version of manuscript.