Knockdown of tyrosine hydroxylase in the nucleus of the solitary tract reduces elevated blood pressure during chronic intermittent hypoxia

Abstract

Noradrenergic A2 neurons in nucleus tractus solitarius (NTS) respond to stressors such as hypoxia. We hypothesize that tyrosine hydroxylase (TH) knockdown in NTS reduces cardiovascular responses to chronic intermittent hypoxia (CIH), a model of the arterial hypoxemia observed during sleep apnea in humans. Adult male Sprague-Dawley rats were implanted with radiotelemetry transmitters and adeno-associated viral constructs with green fluorescent protein (GFP) reporter having either short hairpin RNA (shRNA) for TH or scrambled virus (scRNA) were injected into caudal NTS. Virus-injected rats were exposed to 7 days of CIH (alternating periods of 10% O₂ and of 21% O₂ from 8 AM to 4 PM; from 4 PM to 8 AM rats were exposed to 21% O₂). CIH increased mean arterial pressure (MAP) and heart rate (HR) during the day in both the scRNA (n = 14, P < 0.001 MAP and HR) and shRNA (n = 13, P < 0.001 MAP and HR) groups. During the night, MAP and HR remained elevated in the scRNA rats (P < 0.001 MAP and HR) but not in the shRNA group. TH immunoreactivity and protein were reduced in the shRNA group. FosB/ΔFosB immunoreactivity was decreased in paraventricular nucleus (PVN) of shRNA group (P < 0.001). However, the shRNA group did not show any change in the FosB/ΔFosB immunoreactivity in the rostral ventrolateral medulla. Exposure to CIH increased MAP which persisted beyond the period of exposure to CIH. Knockdown of TH in the NTS reduced this CIH-induced persistent increase in MAP and reduced the transcriptional activation of PVN. This indicates that NTS A2 neurons play a role in the cardiovascular responses to CIH.

sleep apnea is a condition where periodic, repetitive cessation of ventilation results in intermittent nocturnal episodes of arterial hypoxemia. This activates arterial chemoreceptors (CR) leading to increased sympathetic nerve activity (SNA), which contributes to increased arterial pressure (AP) (55). This elevated SNA and AP persist during the daytime when the patients are not experiencing apneic episodes (5). An estimated 15 million Americans suffer from various forms of sleep apnea (67).

The nucleus of the solitary tract (NTS) is the first integrative site of CR input termination within the central nervous system (11, 27), and neurons responding to CR activation are found throughout the NTS (35). However, chemoreceptor afferent integration appears to be more prevalent in the caudal aspects of the NTS, more precisely in the commissural region caudal to the calamus scriptorius (35, 36, 66), an area rich in catecholaminergic neurons, also known as A2 noradrenergic neurons. A2 neurons are activated by chemoreceptor afferents during systemic hypoxia and carotid sinus nerve stimulation (14). A2 neurons mediate responses to a variety of stressors (49) and project to sympathoregulatory sites throughout the neuraxis (43, 49, 65). Among these regions is the paraventricular nucleus (PVN), which receives a major projection from A2 neurons (64), and studies have found that the cardiovascular response to activation of peripheral chemoreceptors involves activation of PVN neurons (42, 48). PVN projections to the rostral ventrolateral medulla (RVLM) have been shown to contribute to the cardiovascular changes occurring during hypoxia exposure in rats (28).

To simulate the hypoxemia that occurs during sleep apnea, a number of labs have chronically exposed rats to intermittent hypoxia (CIH), which results in a persistently elevated AP (18, 29), as seen in humans with sleep apnea. The mechanisms involved in the genesis of hypertension induced by sleep apnea and CIH appear to involve a sequence of events starting with increased arterial chemoreceptor function (16, 30), stimulation, and activation of neurons of central nervous system (CNS) (4, 14, 29, 61) leading to augmented SNA (2, 17). Studies have used accumulation of FosB or its more stable splice variant ΔFosB, a member of activator protein-1 (AP-1) transcription factor family, as a marker for chronic or intermittent activation of CNS neurons (23, 41) following CIH (29). The goal of these studies was to test the hypothesis that depletion of tyrosine hydroxylase (TH) in A2 neurons would attenuate the elevated AP observed during 7 day exposure to CIH by reducing the transcriptional activation of neurons in sympathoregulatory sites like PVN and RVLM.

MATERIALS AND METHODS

Animals.

Experiments were conducted on adult male Sprague-Dawley rats (250–350 g, Charles River Laboratories, Wilmington, MA). Rats were housed in a thermostatically regulated room (23°C) with 12-h light:12 h dark cycle (12 L:12 D, on at 7 AM, off at 7 PM). A 1-wk acclimatization period was provided for the rats before any procedures were performed. Rats were provided with food and water ad libitum. All experimental procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of the University of North Texas Health Science Center.

Telemetry implantation.

Mean arterial pressure (MAP), heart rate (HR), and respiratory frequency (RF) were monitored in conscious rats using a radiotelemetry system (DSI, St. Paul, MN). Under 2% isoflurane inhalation anesthesia and aseptic conditions, all rats were implanted with an abdominal aortic catheter attached to a CA11PA-C40 radiotelemetry transmitter. The transmitter was secured to the abdominal muscle.

NTS microinjections.

Rats were randomly divided into two groups: a group injected with an adeno-associated virus (AAV) with a green fluorescent protein (GFP) reporter and short hairpin RNA for TH (shRNA; n = 15), and a group injected with an AAV with a GFP reporter and scrambled RNA, which is a control for the shRNA virus (scRNA; n = 14). The viral constructs were commercially available (Genedetect) and synthesized at titer 1.0 × 10¹² genomic particles/ml using recently published sequences (63). After 1 wk of recovery from telemetry implantation surgery, under 2% isoflurane inhalation anesthesia, rats were placed in a stereotaxic frame and a limited occipital craniotomy conducted to expose the caudal medulla region. Three 100-nl injections of either shRNA or scRNA were performed with glass micropipette (tip diameter, 50 μm) using a pneumatic picopump (PV 800, WPI, Sarasota, FL) over a 5-min period, at the calamus scriptorius and bilaterally at 0.5 mm rostral and 0.5 mm lateral to calamus to cover the caudal NTS. Because of technical difficulties, telemetry data were recorded from 13 rats in shRNA group.

Chronic intermittent hypoxia exposure.

One week after the AAV injections, rats were transferred to a commercially available hypoxia chamber system where O₂ concentration was varied using 100% N₂, 100% O₂ and a computerized system (Oxycycler, Biospherix, NY). Seven days of baseline MAP, HR, and RF were recorded during which the chambers were maintained at 21% O₂. This was followed by a 7-day exposure to CIH, as described previously (70). Setting the cycles at 9% O₂ for 6 min and 21% O₂ for 4 min resulted in an O₂ concentration of 10% for 3 of the 6 min. The rats were exposed to the CIH during the light phase, which coincides with their sleeping period, from 8 AM to 4 PM (8 h). From 4 PM to 8 AM the rats were exposed to 21% O₂. After the last day of CIH exposure the rats were euthanized and the brain of each rat was collected for either immunohistochemistry or Western blotting.

Immunohistochemistry.

Rats were anesthetized with thiobutabarbital (100 mg/kg ip, Inactin; Sigma, St. Louis, MO) on morning of the day following the last CIH exposure and perfused transcardially with phosphate-buffered saline followed by 4% paraformaldehyde. Brains were then postfixed for 1–2 h before cryoprotecting in 30% sucrose at 4°C. Three sets of coronal 40-μm sections were collected and stored in cryoprotectant at −20°C until processed for immunohistochemistry. Separate sets of serial sections from the brain stem were processed either for TH-dopamine β-hydroxylase (DBH) double labeling or FosB-TH double labeling. Forebrain sections containing the PVN were processed only for FosB. For TH-DBH double labeling, the sections were processed with a primary antibody cocktail of rabbit anti-TH (1:1,000, AB152; Millipore, Billerica, MA) and mouse anti-DBH (1:1,000, MAB308; Millipore) followed by a secondary antibody cocktail with CY3-labeled anti-rabbit (1:250, 711-165-152; Jackson Immunoresearch) and AMCA-labeled anti-mouse (1:100, 715-156-150; Jackson Immunoresearch). For FosB-TH double labeling, a separate set of brain stem sections were first processed for FosB using a goat anti-FosB primary antibody (1:5,000, sc-48-G, Santa Cruz Biotechnology, Santa Cruz, CA) and a biotinylated horse anti-goat IgG (1:200, BA-9500; Vector labs Burlingame, CA). Next the sections were reacted with an avidin-peroxidase conjugate (Vectastain ABC kit, PK-4000; Vector labs) and PBS containing 0.04% 3,3′-diaminobenzidine hydrochloride and 0.04% nickel ammonium sulfate for 10–11 min. After rinsing was completed, these FosB-stained brain stem sections were processed with mouse anti-TH primary antibody(1:1,000, MAB318; Millipore) and a CY3-labeled donkey anti-mouse secondary antibody (1:250, 715–165-150; Jackson ImmunoResearch). Forebrain serial sections were only stained for FosB as described above. The FosB primary antibody used in this study does not discriminate ΔFosB from FosB; therefore the immunoreactivity will be referred to as FosB/ΔFosB.

Imaging and cell counts.

Imaging of TH-DBH immunoreactive neurons and GFP expression in the NTS was performed using an Olympus IX-2 DSU confocal microscope (Olympus, Tokyo, Japan) equipped for epifluorescence. Sections double labeled for TH-DBH were used to count the number of TH and DBH immunoreactive neurons and to construct the images in Figs. 2 and 3. FosB immunolabeled forebrain sections were used to count the number of FosB/ΔFosB immunoreactive neurons in PVN and for images in Fig. 5. Sections double labeled for TH-FosB were used to count FosB/ΔFosB immunoreactivity in RVLM and also to acquire images for Fig. 6 using Olympus microscope (BX41) equipped for epifluorescence, and an Olympus DP70 digital camera with DP manager software (version 2.2.1). ImageJ software (version 1.44, NIH, Bethesda, MD) was used to count the number of TH, DBH (shRNA, n = 9 and scRNA, n = 8), FosB/ΔFosB in PVN, and RVLM (shRNA, n = 7 and scRNA, n = 7) immunoreactive neurons in 9 to 10 sections bilaterally for NTS and RVLM; 5 to 6 sections for PVN [2 to 3 sections per rat/subnuclei dorsal parvocellular (dp), medial parvocellular (mp), lateral parvocellular (lp), and posterior magnocellular (pm)] from each rat and the average number of immunoreactive neurons per section calculated for each rat. The regions of interest were identified using the rat brain stereotaxic atlas (45) and within the regions, the counts were conducted in areas as previously described (29).

Fig. 2.

shRNA selective knockdown of tyrosine hydroxylase (TH) in A2 neurons (A) green fluorescent protein (GFP) staining showing the adeno-associated virus (AAV)-TH-shRNA-infected cells in the nucleus tractus solitarius (NTS). B: red-colored Cy3-labeled A2 neurons expressing TH. C: AMCA-labeled blue-colored dopamine β-hydroxylase (DBH)-positive cells. D: superimposed image of A, B and C. E, F, G, and H are the high-magnification images of the area in box of A, B, C, and D, respectively. Arrows point at the cells where in shRNA successfully knocked down TH without affecting DBH's expression. cc, central canal. Scale bar, 50 μm in low-magnification images and 10 μm in high-magnification images. Sections from subpostremal NTS.

Fig. 3.

Scrambled virus does not affect the expression of TH in A2 neurons. A: GFP staining showing the AAV-Sc virus-infected cells in the NTS. B: red-colored Cy3-labeled A2 neurons expressing TH. C: AMCA-labeled blue-colored DBH positive cells. D: composite image of A, B and C. E, F, G, and H are the high-magnification images of the area in box of A, B, C, and D, respectively. Superimposed image of E, F, and G in H showing mostly purple-colored cells, where in both TH and DBH are intact in scrambled virus-infected cells. cc, central canal. Scale bar, 100 μm in low-magnification images and 10 μm in high-magnification images. Sections from caudal NTS.

Fig. 5.

Representative digital images of FosB/ΔFosB staining on one side in PVN regions of scRNA (A) and shRNA (B) rats. Parvocellular and magnocellular subdivisions are diagrammed in the scRNA (A). Image C indicates the mean no. of FosB/ΔFosB-positive cells counted in the different regions with in paraventricular nucleus (PVN). *P value <0.05 vs. scRNA; n = 8/group. dp, Dorsal parvocellular; mp, medial parvocellular; pm, posterior magnocellular; lp, lateral parvocellular (not shown in image A, as it is found at different level of PVN), Scale bar = 100 μm.

Fig. 6.

Representative digital images of FosB/ΔFosB staining on one side in RVLM region of scRNA (A) and shRNA (B). C and D are the TH images of scRNA and shRNA, respectively, E and F are the colocalized images of FosB/ΔFosB and TH images of scRNA and shRNA, respectively. G: mean no. of FosB/ΔFosB-positive cells counted in the rostral ventrolateral medulla (RVLM). Arrows point at the TH-FosB/ΔFosB colocalization. n = 7/group. Scale bar = 100 μm.

Western blot.

Six rats each from shRNA and scRNA groups were anesthetized with thiobutabarbital (100 mg/kg ip Inactin; Sigma) and decapitated quickly. The brain stem was removed and snap frozen in isopentane on dry ice and placed in a brain matrix (Stoelting, Wood Dale, IL) to cut 1-mm thick slices. Caudal and subpostremal NTS regions were dissected from the frozen sections and subjected to sonication in modified radioimmunoprecipitation buffer supplemented with protease and phosphatase inhibitors followed by centrifugation at 10,000 g at 4°C to obtain a clear supernatant of protein. Bradford assay was conducted to determine the total protein concentration. Fifty micrograms of tissue lysate per sample were loaded onto sodium dodecyl sulfate (SDS)-10% acrylamide gel and electrophorosed before transferring to polyvinylidene fluoride (PVDF) membrane. The membrane was then blocked with 5% (wt/vol) nonfat milk in Tris-buffered saline 0.05% (vol/vol) Tween 20 (TBS-Tween; 50 mM Tris base, 200 mM NaCl, 0.05% Tween 20) followed by overnight incubation with primary antibodies against TH (1:1,000 mouse anti-TH, MAB318; Millipore) and/or (glyceraldehyde-3-phosphate dehydrogenase) GAPDH (1:1,000, mouse anti-GAPDH, MAB374; Millipore). Peroxidase-Conjugated AffiniPure sheep anti-mouse IgG (1:5,000, A5906; Sigma Aldrich) was used as a secondary antibody. Immunoreactive bands were detected by enhanced chemiluminescence (ECL reagents, Amersham, Piscataway, NJ) by acquiring digital gel images using Syngene G-box (Frederick, MD). ImageJ software was used to analyze the densitometry of immunoreactive bands.

Telemetry data analysis.

MAP (sampled at 250 Hz), HR, and RF were recorded for 10 s every 10 min as previously described (29, 70). Pulse interval and fluctuations obtained from the AP waveform were used to calculate HR and RF respectively (Dataquest, DSI, MN). MAP, HR, and RF were averaged for every hour in the 24-h period and the 1-h averages averaged during the light phase (period of exposure) during CIH (8 AM to 4 PM) and during the dark period (7 PM to 7 AM).

Statistical analysis.

All data are presented as means ± SE. Effects of CIH on MAP, HR, and RF during different periods of the day (light with CIH and dark with normoxia) in shRNA and scRNA were determined by two-way ANOVA with repeated measures (SigmaPlot, Systat Software, San Jose, CA). Fisher LSD post hoc test was used to identify significant difference among mean values. One-way ANOVA was used to determine any differences between the control day baseline averages of the three parameters between two groups, differences between the scRNA and shRNA groups TH-immunoreacive (ir) cell counts, DBH cell counts, FosB/ΔFosB immunoreactivity in PVN and RVLM, and Western blot means values. P value <0.05 was considered statistically significant.

RESULTS

Responses to CIH in conscious rats.

Figure 1 shows the average changes from baseline MAP, HR, and RF in the light period (A, C, and E) and the dark period (B, D, and F) in conscious rats during CIH. Seven days of control baseline values were recorded in both groups of rats. The average baseline values during the control period of MAP (mmHg), HR (beats/min), and RF (breaths/min) during light phase and dark phase are presented in Table 1. There was no significant difference between the average control days MAP and RF of shRNA and scRNA groups during light or dark phase. However, HR was greater in shRNA compared with scRNA (P < 0.001) during both the light and dark phases.

Fig. 1.

Graphs are presented as differences from baseline with the baseline absolute values represented as zero. 7 days of chronic intermittent hypoxia (CIH) are presented on x-axis as ih1 to ih7. A: CIH caused a significant elevation in mean arterial pressure (MAP) starting day 2 in both short hairpin RNA (shRNA) and scrambled (scRNA) virus-injected groups, compared with their baseline, during light phase. B: sustained hypertension during normoxic dark phase has been significantly reduced in shRNA virus-injected rats. C: CIH significantly increased the heart rate (HR) in both the virus-injected groups. D: during the dark phase, shRNA moderately decreased the HR elevation noticed in scRNA group. E: light phase rrespirationn frequency (RF) significantly increased in shRNA and scRNA viruses-injected groups. F: This elevation in RF was significantly decreased by shRNA. (shRNA; n = 13 and scRNA; n = 14); *shRNA group significantly different from scRNA group. P value <0.05 was considered significant.

Table 1.

Average baseline values of physiological parameters in shRNA and scRNA groups

	Light Phase		Dark Phase
	scRNA	shRNA	scRNA	shRNA
MAP, mmHg	100.9 ± 3.4	100.9 ± 2.3	104.1 ± 4.2	104.9 ± 2.7
HR, beats/min	311.0 ± 6.1*	320.5 ± 5.4	363.3 ± 7.6*	378.4 ± 6.3
RF, breaths/min	95.4 ± 1.3	95.0 ± 1.5	94.4 ± 4.5	96.4 ± 2.9

Values are means ± SE.

cRNA, scrambled RNA; shRNA, short hairpinRNA; MAP, mean arterial pressure; HR, heart rate; RF, respiration frequency.

^*Significantly different from respective phase HR of shRNA.

The average of MAP absolute values of each CIH day during light phase, compared with the baseline values in the Table 1, were significantly higher on all days of CIH except day 1 in both shRNA and scRNA groups (P < 0.05). In the dark phase of CIH, MAP of all the 7 days were significantly higher in scRNA group but were only significantly elevated on days 2 and 5 in shRNA group compared with their respective baseline values. When the changes in MAP were compared, the shRNA-treated rats showed a significantly reduced CIH-induced increase in MAP compared with scRNA-treated rats during the dark phase (Fig. 1B; P < 0.05).

HR on all the days of CIH was significantly higher compared with the baseline in both scRNA and shRNA groups, during the light phase (P < 0.05). This elevation in HR persisted into the normoxic dark phase on all CIH days except day 1 in the scRNA group but not in the shRNA group. The differences in HR between the groups were not significantly different (P = 0.127) (Fig. 1D).

All the days in the scRNA group and the shRNA group, except day 1, showed a significant increase in RF during light phase on CIH days compared with the baseline. During the dark phase, all the days in scRNA-treated rats showed an increased RF compared with the baseline, whereas only days 2 and 5 showed an increase in the shRNA group. Changes in RF during dark phase of CIH days in the shRNA group were significantly lower compared with the scRNA group (Fig. 1F; P < 0.05).

AAV-TH-shRNA reduces the number of TH-ir NTS neurons.

Figure 2 illustrates that the shRNA construct reduced the number of TH-ir neurons in NTS without altering the DBH immunoreactivity. The arrows in Fig. 2, E–H, point out to the neurons where virus-infected cells showed no TH-ir, but the DBH was intact. The number of TH-ir neurons was reduced by 20% in sections with GFP fluorescence in shRNA group; shRNA 28.6 ± 1.5 cells/section compared with scRNA 35.5 ± 1.4 cells/section (P value: 0.005). The number of DBH-immunoreactive neurons was not different between groups (shRNA: 43.8 ± 2.4 cells/section; scRNA: 45.4 ± 2.4 cells/section). The scrambled RNA treatments did not alter the TH-ir in the virus-infected (GFP-expressing) neurons (Fig. 3). There was no apparent difference in the reduction of TH-ir by the shRNA comparing subpostremal and caudal regions of NTS.

Caudal and subpostremal regions showed a reduced TH protein levels in shRNA group.

Counting TH-ir cells is a binary measure, a cell either possesses TH-ir or it does not, and cannot discern partial reductions in the TH content of individual neurons. To confirm TH knockdown as indicated by counting cells with TH-ir, we performed Western blots. In the shRNA group, TH levels in caudal (P value: 0.005; n = 6 in both groups) and subpostremal (P value: 0.02; n = 5, shRNA, n = 6, scRNA) NTS were reduced by 30% and 10%, respectively, compared with scRNA group (Fig. 4).

Fig. 4.

Western blot analysis shows a significant reduction of TH protein levels in caudal NTS (A) and subpostremal NTS (B) of shRNA virus-injected rats compared with scRNA. TH expression was normalized using GAPDH. *P value <0.05 vs. scRNA. Caudal NTS: n = 6/group; subpostremal: n = 5 in shRNA and n = 6 in scRNA.

shRNA group showed a reduced FosB/ΔFosB staining in different regions of PVN.

There was a significant reduction in the number of FosB/ΔFosB-positive cells in the PVN of the shRNA group compared with the scRNA-injected group (P value: 0.006; shRNA: 14 ± 1; scRNA: 34 ± 6, n = 8 in both groups). Further analysis of different regions of PVN displayed a reduction in FosB/ΔFosB-positive cells in dorsal parvocellular (dp; P value: <0.001; shRNA: 14 ± 1; scRNA: 35 ± 2), medial parvocellular (mp; P value: <0.001; shRNA: 15 ± 1; scRNA: 48 ± 3), and lateral parvocellular (lp; P value: <0.001; shRNA: 15 ± 1; scRNA: 33 ± 2) subnuclei. There was little to no FosB/ΔFosB staining in the posterior magnocellular region in either group (pm; shRNA: 3 ± 1; scRNA: 5 ± 1) with no difference between them (Fig. 5).

No change in FosB/ΔFosB staining between the groups in RVLM.

A moderate increase in the number of FosB/ΔFosB-positive cells in the RVLM of the shRNA group compared with the scRNA group (P value: 0.07; shRNA: 19 ± 3; scRNA: 12 ± 2) was not significant. While the FosB/ΔFosB staining in RVLM was intermingled with TH-positive neurons, the increase in FosB/ΔFosB staining was not associated with significant colocalization of FosB/ΔFosB with TH (shRNA: 1.8 ± 0.3; scRNA: 1.6 ± 0.2) (Fig. 6).

DISCUSSION

Sleep apnea is an increasingly recognized contributor to cardiovascular disease and mortality. Sleep apnea patients have elevated blood pressures during both the day and the night, indicating that the effects of sleep apnea persist beyond the actual exposure to nighttime apneas. The model of CIH used in the present study replicates this aspect of the sleep apnea phenotype. It is interesting to note that CIH induces a persistent increase in MAP in the absence of hypercapnia, which accompanies the hypoxia in sleep apnea patients. In fact, due to the increases in ventilation during intermittent hypoxia (IH) in CIH models, the hypoxia is accompanied by hypocapnia. Previous studies in humans (8, 54, 68) and rodents (15, 62) indicate that IH-induced increases in MAP and/or SNA are independent of end-tidal CO₂, being the same whether the subject is hypocapnic, isocapnic, or hypercapnic. The present study provides new insights that suggest catecholaminergic A2 neurons in the NTS are major contributors to the CIH-induced persistent increase in MAP.

A2 neurons are activated by a variety of stressors, including systemic hypoxia (14). A2 projections to sympathoregulatory sites within the CNS (49) suggesting that A2 neurons play a role in the cardiovascular and sympathetic responses to systemic hypoxia. A 7-day exposure to IH produced no change in TH enzymatic activity or protein level in the brain stem; however, this study did not examine specific areas within the brain stem (20). Chronic hypoxia increased TH expression within the A2 cell group (13, 57), and this increase was abolished after carotid sinus nerve section (56). Therefore, the variations in TH expression observed in A2 group are not solely the direct effect of tissue hypoxia but are dependent on afferent chemoreceptor inputs to NTS. A2 neurons also play a role in the hypothalamic-pituitary-adrenal (HPA) axis reactivity as they are found to express glucocorticoid receptors (19, 22) and project to regions important in regulation of HPA axis function (49). Our group has shown that CIH sensitizes HPA axis reactivity (31), and since glucocorticoids have shown to increase arterial pressure in response to acute stress (51, 52), a hyperreactive stress response may also contribute to the CIH induced increase in MAP. Exposure to CIH has been reported to increase the plasma corticosterone levels (71); however, in that study plasma was collected for measurement of corticosterone at euthanasia so the elevated corticosterone level may reflect enhanced stress reactivity.

In this study, we reduced TH levels in A2 neurons using AAV vector delivery of shRNAs directed toward TH and observed MAP, HR, and RF during exposure to CIH. This approach has been used to reduce TH levels in specific regions of the brain in mice (26) and rats (63). Genetic models are widely used to uncover the molecular mechanisms of a disease, but the use of transgenic and knockout models is limited by developmental effects, genetic compensation, and lack of regional specificity. The use of shRNA to knockdown genes of interest in the whole brain (44) and in specific regions (26, 50, 69) has provided a useful adjunct approach to the use of genetic models. Unlike saporin toxin studies (10, 60) the use of shRNA reduces TH levels and does not kill the A2 neurons. To address concerns about the toxicity of these viral constructs and possible knock down of other genes, a previously used construct (63) at a safe titer (>1 × 10¹² genomic particles/ml) was used. In addition, immunohistochemistry was performed for DBH, which like TH, is also specific for catecholaminergic neurons. Figures 2 and 3 illustrate that the viral constructs are not toxic and do not suppress DBH levels. Uniform sized bands for the housekeeping gene GAPDH in the Western blots in both shRNA and scrambled groups also suggest no nonspecific changes in protein levels.

Our finding of a CIH-induced persistent increase in respiratory frequency was surprising as central catecholaminergic neurons are considered to exert a weak, tonic inhibition of ventilation (34). However, the cited study produced a global reduction in neuronal catecholaminergic content, and more selective and site-specific reductions in neuronal catecholamine content might produce different results. It has been suggested that CIH-induced hypertension might be due to increased respiratory drive to sympathoexcitatory neurons in the brain stem (37).

Previous studies from our group have shown that within the first week of exposure CIH increased MAP during the light phase, when the rats are exposed to the intermittent hypoxia, and the increase in MAP persisted into the normoxic dark phase (7, 24, 29). The results of this study demonstrate that TH knockdown in A2 neurons decreased the persistent increase in MAP and RF during the normoxic dark phase. These results demonstrate that A2 neurons contribute to the cardiovascular and respiratory responses to intermittent hypoxia without altering baseline blood pressure.

Consistent with this finding of a reduced response to CIH, a physiological stress, are the results of a recent study that found that injections of DBH-saporin into caudal NTS reduced the number of catecholaminergic neurons and reduced chronic stress-induced elevations in dark phase blood pressure (10). This study and others have shown increases or no change in baseline blood pressure following lesion of A2 neurons with either DBH-saporin (10, 60) or after silencing of A2 neurons by overexpressing potassium channels (12). Our results indicate no change in baseline blood pressure in shRNA-injected rats. This may be result of a more modest reduction in TH in our study; however, such modest reductions could represent a degree of knock down that might happen physiologically. It could also reflect the fact that the shRNA does not kill the A2 neurons. The contribution of A2 neurons to baseline blood pressure is important to understand and requires further study.

To define potential sites in the CNS that might be involved in the reduction in MAP during normoxic dark phase of CIH following decreased TH levels in NTS, we examined FosB/ΔFosB in sympathoregulatory sites like PVN and RVLM. Increased expression of FosB during CIH is an important finding because FosB/ΔFosB has been linked to neuronal adaptations and plasticity under conditions such as drug addiction, epilepsy, and long-term potentiation (6, 32, 33, 38–41). The present results confirm our previous studies (29) demonstrating increased levels of FosB/ΔFosB in CNS neurons following CIH which may mediate changes in gene expression that alter neuronal function. PVN neurons have been shown to increase levels of FosB/ΔFosB immunoreactivity following exposure to CIH (29). PVN parvocellular neurons project to RVLM (46, 47, 53, 58) and intermediolateral cell column of spinal cord (47, 53, 58, 59), the location of sympathetic preganglionic neurons. The reduction in the number of FosB/ΔFosB immunoreactive neurons in the dorsal parvocellular (dp), medial parvocellular (mp) and lateral parvocellular (lp) (Fig. 5) in the shRNA rats, compared with scRNA rats, indicates that TH knockdown in the NTS reduced the transcriptional activation of PVN, most likely by reducing A2 neuron excitation of PVN neurons (49). This indicates that the sympathetic drive from the PVN (9, 21) might have been decreased, causing the reduction in the CIH-induced persistent increase in MAP during the normoxic dark phase.

The observation that shRNA knockdown of TH in the NTS was not associated with a change in the FosB/ΔFosB immunoreactivity in RVLM (Fig. 6) was not expected as decreased MAP was expected to be associated with decreased transcriptional activation of RVLM. However, A2 neurons do not project directly to the RVLM (3, 25). Optogenetic stimulation of catecholaminergic neurons in RVLM in the rat increases SNA and blood pressure (1). Noncatecholaminergic neurons in RVLM serve a variety of functions (e.g., thermoregulation, respiration, cardiovascular). The reduced activity of PVN parvocellular neurons and their projections to the intermediolateral cell column of spinal cord might be the driving force behind the reduction in the dark phase MAP elevation observed after reductions in TH in the caudal NTS.

Perspectives and Significance

This study shows that it is plausible to knockdown TH using viral vectors without affecting the cell vitality and other protein expression. These results suggest that activation of NTS A2 neurons during CIH mediates the increased sympathetic outflow that has been shown to underlie CIH-induced persistent hypertension. The observations that FosB was not different between treatment groups in the RVLM and that the increase in blood pressure during the light phase was also not different between groups raises the possibility that PVN sympathoexcitatory neurons might have a greater contribution to the dark phase elevation in baseline blood pressure, while the RVLM might be more involved during the light phase exposures to hypoxia.

GRANTS

This study was funded by the National Heart, Lung, and Blood Institute Grant HL-088052.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

Author contributions: C.S.B. and S.W.M. conception and design of research; C.S.B., A.R., and M.F. performed experiments; C.S.B., A.R., K.Y., and J.T.C. analyzed data; C.S.B. and S.W.M. interpreted results of experiments; C.S.B. and A.R. prepared figures; C.S.B. drafted manuscript; C.S.B., J.T.C., and S.W.M. edited and revised manuscript; C.S.B., A.R., M.F., K.Y., J.T.C., and S.W.M. approved final version of manuscript.