Activation of acid-sensing ion channels contributes to the hypercapnia-induced neurovascular response in the neonatal cerebral cortex

Abstract

The cerebral microcirculation undergoes dynamic changes regarding neurovascular responses through the perinatal period. CO₂ is one of the most potent regulator of cerebral blood flow, however its effector mechanisms involve multiple possibly yet unknown pathways in the neonate. Acid-sensing ion channel-1A (ASIC1A) appears to play a decisive role in this cerebrovascular response in adult mice, however, cortical ASIC1A expression and its possible contribution to the response in the translational piglet model of term neonates is unknown. Therefore, we investigated neocortical ASIC1A expression and the effect of the specific ASIC1A-inhibitor psalmotoxin-1 on the neurovascular response to graded hypercapnia in piglets. Anesthetized, mechanically ventilated newborn pigs were equipped with a closed or open cranial window to assess the cortical blood flow (CoBF) with laser speckle contrast imaging or the neuronal activity with multichannel electrodes, respectively. Graded hypercapnia was elicited by ventilation of 5%–10% CO₂. ASIC1A was found to be ubiquitously expressed in neocortical neurons. Psalmotoxin-1 while unaffected the CoBF response to 5% CO₂, it significantly attenuated further increases in CoBF to 10% CO₂, accompanying the highly altered hypercapnia-induced changes in neuronal activity. We conclude that ASIC1A activation by hypercapnia-induced acidosis is partially responsible for the neurovascular response in the neonatal brain.

Introduction

Selective increases in arterial CO₂ levels (hypercapnia) evoke a robust cerebrovascular response marked by vasodilation in virtually all segments of the cerebrovascular tree and a consequent increase in cerebral blood flow. Although this cerebrovascular response is quite conspicuously preserved in all studied mammalian species, there are notable differences in the so far identified mechanisms of CO₂ reactivity, for instance there are differences among different species, and there are also differences within a species dependent on the developmental stage (prenatal, perinatal, juvenile, and adult brain), the studied brain region, or the studied cerebrovascular segment (e.g., cerebral arteries, pial arterioles, or intraparenchymal vessels).^1–5 In term neonates, the cerebrovascular reactivity to CO₂ appears to be already developed at birth that is further maturing in the early postnatal period perhaps as a part of the cardiorespiratory adaptations to the extrauterine environment. ⁶ In newborn pigs, an established model of the term human neonate, hypercapnia-induced increases in regional cerebral blood flow were shown to be larger in the brainstem and subcortical structures than in the cerebral cortex,^7–9 but similar studies on regional blood flow responses in term neonates are lacking despite that regional cerebral blood flow differences clearly exist. ¹⁰ In piglets, the cerebrovascular dilation was studied mostly in pial arterioles on the surface of the cerebral cortex using closed cranial window combined with intravital microscopy.^11–13 From a series of studies, the main mechanism of the pial arteriolar vasodilation to hypercapnia emerged as an endothelium-dependent vasodilation depends mainly on prostanoids shown by the hypercapnia-induced increases in prostanoid levels and the sensitivity of the response to endothelial injury and indomethacin. ¹⁴ However, as low levels of exogenous prostanoids could restore the cerebrovascular response to hypercapnia after abolishment by indomethacin, a more permissive rather than a mediator role of prostanoids has been accepted. In addition, indomethacin likely interferes with other yet unknown mechanisms unrelated to cyclooxygenase as other non-selective cyclooxygenase inhibitors such as ibuprofen or acetaminophen or selective inhibitors of cyclooxygenase-1 or -2 could not reproduce the inhibitory effect of indomethacin on the hypercapnia-induced pial arteriolar response in piglets in a later study. ¹⁴ Furthermore, endothelial-derived carbon monoxide has also been shown to play an active vasodilatory role in the pial arteriolar responses in piglets whereas nitric oxide provided a permissive signal to this carbon monoxide-mediated vasodilation.^15,16 However, developmental maturation of the response assessed in juvenile pigs revealed the increasing contribution of nitric oxide suggesting even a developmental shift in the molecular machinery of the response. ¹⁷

More recently, using laser-speckle flowmetry, we observed that the hypercapnia-induced cortical blood flow (CoBF) response was abolished by prior local administration of NMDA on the cortical surface. ¹⁸ In the piglet model, NMDA was known to elicit CoBF increases in vivo, but isolated piglet pial arterioles do not respond to NMDA ex vivo.^19,20 Our somewhat surprising finding therefore suggests that the mechanism of the cerebrovascular response to hypercapnia in piglets must involve important contribution from parenchymal/neural elements that is disrupted by the NMDA stimulus.

Faraci et al. ²¹ reported that hypercapnia-induced pial arteriolar vasodilation was critically dependent on the availability of functional acid-sensing ion channels (ASIC1A) as either selective neuronal ASIC1 knockout or pharmacological inhibition by the highly selective inhibitor psalmotoxin-1 (PcTx1) virtually abolished the response in adult mice. ²¹ Therefore, we hypothesized that ASIC-s can perhaps be also involved in a neurovascular mechanism responsible for the CoBF response to graded hypercapnia in the neonate as well. To test this hypothesis, after demonstrating ASIC1A expression in the neonatal cerebral cortex, we assessed the effect of PcTx1 on both the cortical neuronal and the vascular response to graded hypercapnia in newborn pigs. As PcTx1 affected the cerebrovascular response, we also assessed if amiloride, an inhibitor of Na⁺ channels similar to ASIC1A but not a potent inhibitor of ASIC1A itself would affect the cerebrovascular responses to graded hypercapnia.

Materials and methods

Animals and surgery

Newborn male Landrace piglets (n = 21, body weight: 1.7 ± 0.2 kg) were obtained from Pigmark Ltd., Co., Szeged, Hungary. The experimental procedures were approved by the National Ethical Committee on Animal Experiments (ÁTET, I.74–7/2015) and the permit was issued by the National Food Chain Safety and Animal Health Directorate of Csongrád county, Hungary (permit nr: XIV./1414/2015). All procedures were performed according to the guidelines of the Scientific Committee of Animal Experimentation of the Hungarian Academy of Sciences (updated Law and Regulations on Animal Protection: 40/2013. (II. 14.) Gov. of Hungary), following the EU Directive 2010/63/EU and the ARRIVE guidelines.

For the laser speckle contrast imaging (LSCI)/electrophysiology experiments (n = 18), the animals were anesthetized with intraperitoneal sodium thiopental injection (45 mg/kg; Sandoz, Kundl, Austria) and their core temperature was maintained at 38.5 ± 0.5°C with a servo-controlled heating pad (Blanketrol III, Cincinnati SUB-zero, Cincinnati, OH, USA). The piglets were mechanically ventilated with humidified air (FiO₂ = 0.21) with the following ventilation parameters: peak inspiratory pressure (PIP): 12–14 cm H₂O, respiration rate (RR): 30–32/min. Anesthesia/analgesia was maintained via an intravenous catheter using morphine (100 μg/kg bolus then 10 μg/kg/h; Teva, Petach Tikva, Israel) and midazolam (250 μg/kg bolus then 250 μg/kg/h; Torrex Pharma, Vienna, Austria) as used in our previous studies^1,2 along with supportive fluid therapy (0.45% NaCl, 5% glucose; 3 ml/kg/h). Monitoring the mean arterial blood pressure (MABP), heart rate (HR), and taking blood samples were assessed via a catheter inserted into the carotid artery. Blood samples were analyzed for pH, gases, metabolites and electrolytes using the epoc^® Blood Analysis System (Epocal Inc., Ottawa, ON, Canada). The peripheral saturation was monitored using pulse oximetry (SpO₂).

The heads of the animals were fixed into a stainless steel stereotaxic frame (RWD Life Science, Shenzhen, Guangdong Province, China). Stainless steel closed cranial window was implanted over the parietal cortex for the LSCI studies which was sealed with bone wax and cemented with dental acrylic. ²² For the electrophysiology measurements we used an open cranial window at the same location (Figure 1(a)). The reference and ground electrodes were mounted to the frontal bone. We carefully removed the dura mater for both measurement types, avoiding the blood vessels, and if necessary the smaller veins were cauterized. The subarachnoidal space was filled with warmed (38°C) artificial cerebrospinal fluid (aCSF) containing 7.71 g/l NaCl, 0.22 g/l KCl, 0.221 g/l CaCl₂, 0.132 g/l MgCl₂, 0.665 g/l dextrose, 0.402 g/l urea, and 2.066 g/l NaHCO₃, and was equilibrated with a gas mixture containing 6.3% O₂, 6.2% CO₂, and 87.5% N₂. At the end of the experiments the animals were euthanized with an overdose of pentobarbital sodium (300 mg/kg, Release; Wirtschaftsgenossenschaft deutscher Tierärzte eG, Garbsen, Germany).

Figure 1.

Experimental protocol. (a) The animals were fitted with either closed or open cranial window for LSCI (n = 11) and electrophysiology (n = 7), respectively. (b) In all animals, graded hypercapnia was elicited three times by ventilating 5% and 10% CO₂ for 7–7 min. The cranial window was continuously perfused with artificial cerebrospinal fluid (aCSF, 0.5 ml/min). Starting 30 min before and ending 15 min after the second graded hypercapnia, amiloride, or psalmotoxin-1 (PcTx1), a selective inhibitor of the acid-sensitive ion channel 1A obtained from the venom of the tarantula (Psalmopoeus cambridgei), dissolved in the aCSF (10 nM) was applied topically onto the cortex.

For the detection of cerebrocortical expression of ASIC1A in naïve animals (n = 3), after anesthesia with intraperitoneal sodium thiopental injection (45 mg/kg), the brains were harvested and tissue samples have been prepared for further analysis as described. ²³

Experimental protocol of repeated graded hypercapnia and PcTx1/amiloride application

The animals (n = 18) were divided into three groups for the LSCI (n = 11) and the electrophysiological (n = 7) studies, respectively. In all three groups we used the same experimental protocol: after obtaining the baseline recordings, graded hypercapnia was induced by ventilation with 5% and 10% CO₂ for 7 min for each concentration. Hypercapnia was followed by topical application of PcTx1 or amiloride (Merck KGaA, Darmstadt, Germany), dissolved in aCSF and perfused through the cranial window for 30 min (10 nM, 0.5 ml/min; 10 µM, 0.5 ml/min respectively), special care was taken to protect the light-sensitive PcTx1 from ambient light. Then, the second graded hypercapnia was elicited while PcTx1/amiloride perfusion continued until 15 min after the completion of hypercapnia. The PcTx1/amiloride was then removed by perfusion of the cranial window with aCSF for 20 min that was followed by repeating the graded hypercapnia for the third time (Figure 1(b)). The doses of PcTx1 and amiloride were selected based on the previous literature.^21,24,25

LSCI analysis

The brain was illuminated with polarized light (λ = 808 nm, 200 mW) with a laser diode (DL-8141-002 SANYO Electric Co., Japan) and the images were recorded with a monochrome camera (PL-B741F; PixeLINK^®, Ottawa, ON, Canada). The speckle images were sampled at 1 Hz with 1 ms exposure time during all stimuli. The LSCI analysis was performed offline in LabView (National Instruments Co., Austin, TX, USA). The contrast maps were calculated from the raw speckle images. The τ correlations were calculated using equation (1) where T is the exposure time, K(T) is the speckle contrast and β is the coherence factor. All values were expressed as percentages of baseline. Four different region of interests (ROI) were selected in each animal for both parenchymal flow (5 × 5 pixels, ~1200 µm²) and pial arteriolar diameter (130 × 160 pixels) measurement where peak CoBF and diameter changes were observed. The values from the four ROIs were averaged in each animal.

$$K\left(T\right)=\sqrt{\mathrm{\beta }}{\left\{\frac{{\tau }^2}{2T^2}\left[exp\left(\frac{-2T}{\tau }\right)-1+\frac{2T}{\tau }\right]\right\}}^{\frac{1}{2}}$$ (1)

To obtain a better resolution for pial arteriolar diameter measurements, 30 images were averaged and Otsu filtering method was applied. To acquire the vessel diameters we used Canny image segmentation and Euclidean distance measurement in MATLAB (Mathworks Inc., Natick, MA, USA).^18,26

Electrophysiological recordings

All recordings were taken with 16-channel, acute single shank silicone probes (A1x16-10 mm-100-177-A16; Neuronexus Technologies. AnnArbor, MI, USA). Data acquisition was done with RHD2000 electrophysiological recording system (Intan Technologies, Los Angeles, CA, USA) under a Faraday cage. Broad band data was sampled at 20 kHz, for the local field potentials (LFP) we downsampled the data to 1250 Hz along with notch (50 Hz) and lowpass Butterworth (400 Hz) filtering. All data were analyzed in MATLAB environment with the appropriate toolboxes (Signal Processing and Image Processing toolbox) and custom written scripts. For the LFP spectral analysis, we decomposed the signal to the physiological frequency ranges—delta (1–4 Hz), theta (4–8 Hz), alpha (8–13 Hz), beta (13–30 Hz)—for the step by step (1 s) calculation of the power spectral density (PSD) with Fast Fourier Transformation (FFT), using Gaussian window (30 s). Spike sorting was performed with the Klusta package, using automatic spike detection with Spikedetekt2 and clustering with KlustaKwik2. ²⁷ We used Phy KwikGUI (https://github.com/cortex-lab/phy) for the manual adjustment of the clusters. The noise and multi-unit clusters were discarded. The putative pyramidal cells and interneurons were identified by their autocorrelograms (ACG) and spike waveform characteristics.

Total RNA extraction

Total RNA was extracted from three regions of the cerebral cortex (sigmoid gyrus, coronal gyrus, and the occipital pole) with Tri Reagent (Sigma, MA, USA) according to the manufacturer’s protocol. Briefly, 1 ml Tri Reagent was added to 50–100 mg tissue samples and after homogenization by Dounce tissue grinder and incubation 5 min at room temperature to allow dissociation of nucleoprotein complexes, 0.2 ml of chloroform was added. The samples were mixed vigorously and then centrifuged at 12,000 × g for 15 min at 4°C. After centrifugation, the RNA was precipitated from the upper, colorless aqueous phase with 0.5 ml of isopropanol. The samples were then incubated at room temperature for 10 min and centrifuged at 12,000 × g for 10 min at 4°C. The supernatant was removed and the RNA pellet was washed once with 70% ethanol and centrifuged at 7500 × g for 5 min. The pellet was air dried and dissolved in diethyl pyrocarbonate (DEPC)-treated water. Total RNA quantity (OD260) and purity (OD260/280) were measured by a NanoDrop spectrophotometer (Thermo Scientific, Waltham, MA, USA); all samples that were used for further analysis exhibited an absorbance ratio in the range of 1.6–2.0.

cDNA synthesis and quantitative PCR analysis

For quantitative PCR (qPCR) 1 μg of total RNA was reverse transcribed using the iScript cDNA synthesis kit (Bio-Rad, cat.no. 1708891) according to the manufacturer’s protocol with random hexamer priming. qPCR was performed in a thermo cycler (Bio-Rad CFX96^TM Optics Module) real-time system with the iQ™ SYBR^® Green Supermix (Bio-Rad, cat.no. 1708882) and primers for ASIC1 and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) were designed specifically using the Sus scrofa domestica (white pig) National Center of Biotechnology Information (NCBI) reference sequence database (https://www.ncbi.nlm.nih.gov/Entrez) as follows:

• ASIC1: Fw (5’-3’): CGACATTCGAGACATGCTGC, Rev (5’-3’): TTCCATACCGCGTGAAGACC; (Ref. seq.: XM_021091639.1)

• GAPDH: Fw (5’-3’): GTGAACGGATTTGGCCGC, Rev (5’-3’): AAGGGGTCATTGATGGCGAC; (Ref. seq.: NM_001206359.1)

The thermal cycling condition included an initial denaturation step at 95°C for 30 s and 40 cycles of denaturation at 95°C for 10 s, annealing at 59°C for 30 s, and extension at 72°C for 20 s.

To check the amplification specificity, the qPCR was followed by a melting curve analysis. The reference gene GAPDH was chosen to standardize the quantitative experiments, and a no-template was used as negative control where RNAse-free water was added instead of cDNA. For each triplicate sample threshold cycles (Ct) were calculated for ASIC1 and GAPDH genes, and the normalized gene expressions were calculated by the ΔΔCt method. ²⁸ Statistical comparison of qPCR data was performed by comparing the ΔΔCt values.

Western blotting

Brain sections were removed and flash-frozen in liquid nitrogen and stored at −80°C until use. The tissue samples were homogenized in ice-cold RIPA buffer (Millipore, Cat.no. 20-188) by Dounce tissue grinder. The homogenate was centrifuged at 12,000 × g for 20 min at 4°C and the supernatant was collected. The protein concentration was determined using Pierce^TM BCA protein assay kit (Thermo Scientific™, Cat.no. 23225). The samples with protein concentration adjusted to 30 µg were subjected to polyacrylamide (10%) gel electrophoresis for separation. Proteins were transferred to a 0.45 μm pore-sized nitrocellulose blotting membrane (Amersham^TM Protran^TM, Lot.no. A29642915). The membrane was rinsed with Tris-buffered saline mixed with Tween-20 (0.1%; TBST) and incubated in 1% bovine serum albumin (BSA; Sigma, Cat.no. A7030) or milk powder (Serva, Lot.no. 200120). After washing in TBST 3 times for 5 min, the membrane was incubated overnight at 4°C with anti-ASIC1 antibody (Genetex, Cat.no. GTX80451) diluted in TBST with 0.1% milk powder and anti-β-actin antibody (Abcam, ab20272) diluted in TBST with 0.1% BSA. Then a 3 min × 10 min washing procedure was followed by incubation for 2 h at room temperature with peroxidase-conjugated antibodies: polyclonal goat anti-rabbit and polyclonal rabbit anti-mouse antibodies, respectively (Dako, Ref. nos. P0448 and P0161). The protein-bound antibodies were detected using Pierce^TM ECL Western Blotting Substrate (ThermoScientific^TM, Lot.no. TF269477) and visualized using Li-Cor Odyssey XF imaging system (LiCorbio^TM). The protein levels were analyzed by the Quantity One software version 4.6.6 by normalizing to β-actin.

Fluorescent immunohistochemistry

For immunofluorescence studies, ASIC1A-HuC/HuD double labelling immunohistochemistry were performed on brain paraffin sections (5 µm).^23,29 After blocking in tris(hydroxymethyl)aminomethane-buffered saline (TBS) containing 1% bovine serum albumin and 10% normal goat serum, the sections were incubated overnight at 4°C with anti-ASIC1 (rabbit polyclonal IgG, 1:50; Cat. No. GTX80451, GeneTex, Irvine, CA, USA) and pan-neuronal anti-HuC/HuD (mouse monoclonal IgG, 1:50; Cat. No. A-21271, Invitrogen, USA) primary antibodies. After washing in TBS with 0.025% Triton X-100, sections were incubated with anti-rabbit Alexa Fluor 488 (1:200; Cat. No. A11008, Invitrogen, Thermo Fisher Scientific, USA) and anti-mouse Cy^TM3 (1:200; Cat. No. 115-165-003, Jackson ImmunoResearch Laboratories, USA) secondary antibodies for 1 h at room temperature. Negative controls were performed by omitting the primary antibodies when no immunoreactivity was observed. All immunostained sections were counterstained with DAPI and mounted with Fluoromount^TM Aqueous Mounting Medium (Sigma-Aldrich, Hungary). Representative photographs were made with a Zeiss Imager Z.2 fluorescent microscope equipped with an Axiocam 506 mono camera (Zeiss, Germany).

Statistical analysis

Statistical analysis was performed in IBM SPSS Statistics 22. All data were tested for normal distribution, using Shapiro–Wilk’s test. To determine the relative CoBF and arteriolar changes we used univariate ANOVA with repeated measures, followed by Tukey’s post hoc test (Supplemental Table 1). For the relative PSD changes we applied multivariate ANOVA (MANOVA) with repeated measures, using Tukey’s post hoc test as well (Supplemental Table 2). p < 0.05* was considered as significant. All results show mean ± SD, respective to the baseline. The Z-score computations were performed in MATLAB and the ΔZ ⩾ 2* was determined as significant.

Results

ASIC1A is expressed in the neocortical neurons of newborn pigs

Using qPCR, we found that ASIC1A mRNA levels were detectable and similar in the frontal, parietal and occipital cortices of piglets (Figure 2(a)). The similar mRNA levels were also reflected in similar amounts of ASIC1A protein levels assessed by Western blot analysis for all 3 examined neocortical areas (Figure 2(b), Supplemental Figure 1).

Figure 2.

ASIC1A expression in the piglet cerebral cortex. (a) qPCR shows similar relative expression of ASIC1A in all assessed cortical regions (n = 3-3, mean ± SD). (b) Western blot analysis revealed the predicted molecular bands for ASIC1A (55 kDa) and β-actin (42 kDa), the ASIC1A/β-actin ratios were similar in all cortical regions. (c) Representative photomicrographs at two magnifications (upper and lower 4-4 images) showing immunofluorescence of ASIC1A (green), the neuronal marker HuC/HuD (red), the nuclear stain DAPI (blue), and the merged triple-labeled image in the piglet parietal cortex. Many cortical neurons in virtually all cortical layers show ASIC1A immunofluorescence. The lower panel of high-magnification images are taken from layer V.

Fr.: frontal; Occ.: occipital; Par.: parietal cortices, HuC/HuD-HuCD.

Immunohistochemistry was employed to reveal the cellular localization of ASIC1A in the piglet cerebral cortex. Prominent neuronal but not glial anti-ASIC1A immunoreactivity was observed in virtually all cortical layers with quite homogenous and similar distribution among the different cortical areas (Figure 2(c)). The neuronal ASIC1A immunoreactivity appeared to have mainly localized to the neuronal soma and the primary dendrites. In addition to the prominent neuronal staining, some vascular immunostaining was also discernible.

PcTx1 does not alter graded hypercapnia-induced changes in physiological parameters

The assessed physiological parameters were all in their respective physiological ranges at the beginning of the LSCI/electrophysiological experiments (Table 1). The normoxic graded hypercapnia, that was evoked three times, resulted in the expected changes in the determined blood gases, metabolites, and cardiovascular parameters during all three stimulations. Accordingly, graded hypercapnia resulted in graded increases in pCO₂ and drops in pH, while pO₂, SpO₂, and lactate levels remained unchanged confirming that tissue oxygenation remained adequate. The MABP and HR changes in response to the graded hypercapnia reflect the activation of chemoreceptors. Importantly, the values of physiological parameters obtained during all three stimulations were virtually identical, therefore, the topical application of PcTx1 before, during, and immediately after the second graded hypercapnia did not affect the response compared to the first stimulation (Table 1).

Table 1.

Mean arterial blood pressure (MABP), heart rate (HR), arterial oxygen saturation (SpO₂), pH, pO₂, pCO₂, HCO₃⁻, blood base excess (BE(b)), glucose, and lactate values during the three consecutive graded hypercapnic stimulations. The responses to hypercapnia were similar in all three cases, the applied ASIC1A inhibitor PcTx1 did not affect the hypercapnia-induced changes in the assessed parameters.

Stimuli	MABP (mmHg)	HR (min−1)	SpO2 (%)	pH	pCO2 (mmHg)	pO2 (mmHg)	HCO3− (mmol/l)	BE(b) (mmol/l)	Glucose (mmol/l)	Lactate (mmol/l)
Baseline	63 ± 9	135 ± 16	95 ± 1	7.47 ± 0.08	41 ± 9	74 ± 9	28.9 ± 3.5	4.8 ± 3.6	5.3 ± 1.2	1.3 ± 0.4
5% CO2	77 ± 11	147 ± 19	94 ± 2	7.31 ± 0.04	63 ± 8	73 ± 11	31.9 ± 2.8	4.6 ± 2.5	5.6 ± 1.2	1.0 ± 0.3
10% CO2	88 ± 15	166 ± 29	93 ± 2	7.17 ± 0.03	93 ± 11	76 ± 10	33.7 ± 3.7	3.4 ± 3.1	6.0 ± 1.2	0.9 ± 0.3
PcTx1	57 ± 7	135 ± 16	95 ± 1	7.45 ± 0.06	42 ± 8	74 ± 7	29.0 ± 5.5	4.5 ± 5.5	5.3 ± 1.2	1.1 ± 0.3
5% CO2	67 ± 12	140 ± 14	94 ± 3	7.29 ± 0.05	64 ± 9	69 ± 9	31.6 ± 4.1	4.0 ± 3.6	5.5 ± 1.1	0.9 ± 0.3
10% CO2	79 ± 19	165 ± 25	93 ± 3	7.16 ± 0.04	94 ± 12	73 ± 12	33.6 ± 3.4	3.1 ± 2.8	6.2 ± 1.0	0.8 ± 0.3
Wash	57 ± 7	140 ± 18	95 ± 1	7.42 ± 0.06	44 ± 9	70 ± 10	28.5 ± 4.1	3.5 ± 3.6	5.4 ± 1.3	0.9 ± 0.4
5% CO2	67 ± 7	141 ± 20	93 ± 2	7.28 ± 0.04	62 ± 19	66 ± 10	32.1 ± 3.7	4.4 ± 3.4	5.9 ± 0.9	0.9 ± 0.3
10% CO2	76 ± 13	164 ± 36	92 ± 3	7.14 ± 0.07	100 ± 19	72 ± 13	33.6 ± 3.3	2.7 ± 3.5	6.5 ± 1.5	1.1 ± 1.1

PcTx1 differentially affects microvascular responses to graded hypercapnia

LSCI analysis provides two-dimensional cortical perfusion images (Figure 3(a)) suitable to assess semi-quantitative changes in CoBF using regions of interests excluding areas of pial vessels (Figure 3(b)), but also to determine the internal diameter of the pial arterioles (Figure 3(c)).

Figure 3.

Effect of the ASIC1A inhibitor psalmotoxin-1 (PcTx1) on increases in cortical blood flow (CoBF) and pial arteriolar diameters changes to graded hypercapnia. (a) Representative LSCI contrast images obtained before, during, and after repeated graded hypercapnia stimulations. Increasing CoBF during hypercapnia is indicated by decreasing speckle contrast values over parenchymal areas visualized by warmer colors according to the color scale on the right. Simultaneously, the pial arteriolar dilation to hypercapnia is discernible. (b) Peak CoBF increases to graded hypercapnia are expressed as percent of baseline. Local PcTx1 treatment did not affect the CoBF response to 5% CO₂ but significantly decreased the response to 10% CO₂ by approximately 1/3. Removal of the toxin at least partially restored the CoBF response to the higher CO₂ stimulation. (c) 10% CO₂ also resulted in marked pial arteriolar vasodilation but this was unaffected by PcTx1 application. Interestingly, removal of the toxin, however, significantly enhanced the pial arteriolar vasodilation during the third hypercapnia stimulus.

aCSF: artificial cerebrospinal fluid.

*p < 0.05.

The first graded hypercapnia resulted in the expected concentration-dependent, reversible increases in CoBF (Figure 3(b)). ASIC1A inhibition by PcTx1 did not per se affect CoBF, after 30 min PcTx1 application before the onset of the second graded hypercapnia, CoBF was 99 ± 5% of the first baseline. PcTx1 also did not significantly affect the CoBF response to 5% CO₂, the peak increases were 32 ± 5% versus 27 ± 7%. In contrast, the CoBF increases to 10% CO₂ were significantly attenuated by PcTx1: peak CoBF increases were reduced from 63 ± 14% to 41 ± 12%* (*p = 0.016). Upon removal of PcTx1, during the third graded hypercapnia stimulation the peak CoBF increase to 5% CO₂ was still intact (29 ± 10%), and the response to 10% CO₂ was not statistically different from either the first or the second response (48 ± 14%; Figure 3(b)).

We also determined pial arteriolar diameters at baseline and during the peak CoBF responses to 10% CO₂ in all three graded hypercapnia stimulations of 4-4 vessels from each animal (Figure 3(c)). We found that pial arterioles with a baseline diameter > 200 µm were only very weakly dilating to hypercapnia so these vessels (n = 8) were omitted from further analysis shown in Figure 3(c). The baseline arteriolar diameters of the analyzed vessels were thus 146 ± 31 µm (n = 16). PcTx1 did not significantly affect baseline diameters or hypercapnia-induced vasodilation significantly, peak increases were 53 ± 32% versus 42 ± 47% compared to baseline diameters. Interestingly, however, removal of the toxin significantly enhanced the pial arteriolar dilation, peak increases were 104 ± 74%* (*p = 0.033, different from either the 1st or the 2nd hypercapnia).

PcTx1 alters electrophysiological changes induced by graded hypercapnia

PSD analysis confirmed that delta (δ) and theta (θ) activity dominated the recorded LFP under baseline conditions similar to previous studies in this model. ¹⁸ Induction of graded hypercapnia with 5% CO₂ resulted in a transient elevation of spectral power that was most prominent for δ in the most superficial layers (100–300 µm), while for θ both in the superficial (100–400 µm) and the deep (900–1200 µm) layers (Figures 4, 6(a), and 6(b)). The elevation was followed by a PSD depression in both ranges that was further augmented by increasing the inhaled CO₂ concentration to 10%. This depression characteristically originated in the deepest cortical layers and was gradually shifting upward. Upon restoring normocapnia, the PSD depression reversed following an opposite direction (gradual downward shift). PcTx1 did not significantly affect the PSD either in the δ or the θ range. However, PcTx1 abolished the transient θ PSD elevation but the δ and θ PSD depressions persisted. After removal of the toxin, PSD decreased and the third induction of hypercapnia caused a long lasting PSD depression in both bands (Figures 4, 6(a), and 6(b)). The described changes were statistically highly significant, for details see (Supplemental Table 2).

Figure 4.

Representative heat map images of relative (a) delta and (b) theta power spectral densities (PSD) obtained from the normalized PSD data of an experiment. The most striking and consistent feature for both the delta and theta waves was the intracortical development of PSD depression during hypercapnia (the red arrows indicate the direction of PSD depression spreading from the deep layers pointing upward) and the reversal of the depression upon restoration of normocapnia (red arrows pointing downward). At the initiation of hypercapnia with 5% CO₂ transiently increased the delta activity in the superficial layers, and this response was unaltered by PcTx1 (a). However, the even more prominent transient theta activity enhancement was completely diminished by PcTx1, and this response wasn’t restored after the toxin removal (b). Contour plot of the average (n = 7) Z-scores of (c) δ and (d) θ activity. 10% CO₂ application resulted in significant (ΔZ ⩾ 2) decrease in both δ and θ bands. The δ activity change was mostly unaffected by the venom, while the θ shows diminished values under the PcTx1 administration and became further reduced by removing the toxin.

Figure 6.

Spider plots of power spectral densities (PSD) and total as well as layer-specific spiking activity in the piglet cortex over the course of repeated hypercapnia stimulations. Mean (a) delta and (b) theta PSD changes expressed as percent of baseline derived from local field potentials of all channels. (a) 5% CO₂ elicited similar increases in the delta PSD (147.7 ± 24.7, 116.8 ± 30.3, 118.4 ± 31.2) in all 3 cases while 10% CO₂ resulted in a similar reduction in PSD (76.6 ± 28.7, 63.7 ± 10.5, 70.3 ± 20.2). (b) Regarding theta activity, the first 5% CO₂ stimulus slightly increased the PSD (125.6 ± 31.9), whereas in the presence of PcTx1 or after PcTx1 washout the response was not significantly altered (103.3 ± 33.3, 81.5 ± 28.5). In contrast, 10% CO₂ resulted in a consistent, prominent and significant theta PSD depression that was unaffected by PcTx1 or its washout (34.5 ± 11.3, 39.4 ± 14.2, 37.9 ± 12.4). (c) Total spiking activity of n=77 recorded neurons from all cortical layers was largely unaffected by PcTx1 or its washout at baseline and in response to 5% CO₂ (657, 496, 488 spike/min). However, the response to 10% CO₂ was modulated by PcTx1. The first 10% CO₂ stimulus mainly increased neuronal firing in layer IV (e) (943 spike/min). However, in the presence of PcTx1, the increased firing to 10% CO₂ spread to both the more superficial layer I-II/III (d) and the deeper layer V (f) resulting in a higher total firing (1438 spike/min) (e). Interestingly, the PcTx1 washout resulted in a suppression of neuronal spiking response to 10% CO₂ in all layers, affecting the total spiking as well (280 spike/min).

Spike sorting, clustering, ACGs and spike waveform characteristics helped identifying altogether 76 firing neurons, the majority of which were pyramidal cells (n = 63, 83%) and some were interneurons (n = 13, 17%; Figure 5(a)). The action potentials from pyramidal cells could be detected in all layers of the cortex, while the activity from the few interneurons was restricted to layers II/III and IV.

Figure 5.

Effect of PcTx1 on changes in cortical neuronal firing during graded hypercapnia challenges in the cortex. (a) Spike waveforms and corresponding autocorrelograms identify firing pyramidal neurons (red) and interneurons (blue). (b) The total of 76 neurons were identified in the n = 7 experiments, the activity from all individual neurons is shown on the raster plot for the whole experiment of the three repeated hypercapnia challenges. The vertical arrangement of the cell recordings in the raster plot reflect the neuron’s position in the cortex according to their approximate position in the cortical layers shown on the right. The corresponding cortical depths are: I. 100–200 µm, II/III. 200–600 µm, IV. 600–900 µm, V. 900–1500 µm, VI. below 1500 µm. The combined firing rate (stimulus time histogram) is plotted above the raster plot. The red arrows indicate the direction of the cortical excitation/depression (pointing upward) and restoration (pointing downward). PcTx1 greatly augmented the spiking activity of neurons responding to 10% CO₂, however, removal of the toxin resulted in a long lasting depression inactivating approximately half of previously responding neurons, especially in the more superficial cortical layers.

During the repeated graded hypercapnia stimulations there was a marked difference between the responses to the lower grade 5% CO₂ and the higher grade 10% CO₂ challenges. The total number of firing neurons were similar during all three hypercapnia stimulation with 5% CO₂, the number of active neurons were 58, 49, and 53 for 1st, 2nd, and 3rd stimulations, respectively. In addition, the changes in the neuronal firing were also similar regarding the number of neurons decreasing or increasing their activity in all three (657, 496, 488 spike/min respectively) responses (Figure 6(c)–(f)). In contrast, the response to 10% CO₂ was much more affected by PcTx1 treatment. Although the total number of active neurons were similar during the first (n = 63) and second hypercapnia (n = 66), but it reduced to only 34 active neurons during the third stimulation with 10% CO₂. In addition, compared to the first severe hypercapnia, the PcTx1 treatment before/during the second application of 10% CO₂ resulted in substantially increased firing of the neurons (from 943 to 1438 spike/min). Furthermore, we observed a layer specific activation of the cortex. Before PcTx1, the activity increased mainly in layer IV (463 spike/min, Figure 6(e)), while during toxin application layer IV activity was unchanged (486 spike/min, Figure 6(d)) and layer I-II/III (from 259 to 543 spike/min, Figure 6(f)) and layer V (from 222 to 409 spike/min) doubled the firing rate in response to 10% CO₂. Interestingly, PcTx1 removal resulted in an extended inhibition of half of the active neurons eliciting almost complete silence of the neurons in the cortical layers II/III and IV (Figure 5(b), 6(d), and 6(e)).

Amiloride does not affect microvascular responses to graded hypercapnia

Physiological parameters were similar to PcTx1-treated animals at baseline and during repeated hypercapnic challenges (data not shown). Amiloride did not affect the monitored physiological parameters (MABP: 55 ± 5 mmHg, HR: 140 ± 29 min⁻¹, SpO₂: 97 ± 3%, pH: 7.46 ± 0.07, pCO₂: 43 ± 10 mmHg, pO₂: 65 ± 7 mmHg, HCO₃⁻: 31.1 ± 4.3 mmol/l, BE(b): 6.4 ± 4.1, Glucose: 6.0 ± 1.4 mmol/l, Lactate: 1.1 ± 0.4 mmol/l) or the CoBF response to hypercapnia. CoBF increases to 5% and 10% CO₂ were 145 ± 16% and 174 ± 41%, respectively, and the response was unchanged in the presence (144 ± 9% and 163 ± 38%) and after the washout of amiloride (131 ± 17% and 168 ± 49%). Amiloride application did not alter the vasodilation of the vessels, the percent changes in arteriolar diameters were similar to each 10% CO₂ (145 ± 40, 146 ± 32, 150 ± 32%).

Discussion

There are two major new findings in our present study: (1) ASIC1A is widely expressed in the neocortex of newborn pigs, and the ASIC1A immunofluorescence is prominently found in the neocortical neurons. (2) PcTx1, a selective inhibitor of ASIC1A partially inhibits the CoBF response to graded hypercapnia simultaneously affecting the hypercapnia-induced changes in the cortical electrical activity including changes in LFP and neuronal firing. The contribution of ASIC1A activation to hypercapnia-induced neuronal-vascular changes in the neonatal brain are discussed in the following paragraphs.

ASICs are voltage-independent, H⁺-gated cation channels that belong to the epithelial Na⁺ channel (ENaC)/degenerin family, which includes the well-known amiloride-sensitive ENaC channels expressed for instance in the renal tubules. ³⁰ Functional ASICs are formed by homo- or heterotrimeric assemblies of ASIC subunits each containing two transmembrane domains connected by a cysteine-rich extracellular loop.^31,32 In central nervous system neurons, ASICs are mainly formed by ASIC1A, ASIC2A, and ASIC2B subunits, among these ASIC1A being the most commonly expressed subunit, with an activation threshold just below pH = 7.0. ³³ Importantly, ASIC1A and heterotrimeric ASIC1A/2A channels are permeable to both Na⁺ and Ca²⁺ ions and thus their activation can contribute to elevations in intracellular Ca²⁺ levels of neurons.^34,35 ASICs are known to be involved in many neurological functions, such as pain, fear, learning, and memory.^36,37 In addition, ASIC-mediated depolarization and Ca²⁺ overload has been found to contribute to hypoxic/ischemic neuronal damage identifying ASIC1A inhibition as a target of neuroprotective strategies. ³⁸

In the newborn pig, a commonly employed large-animal model of neonatal hypoxic-ischemic encephalopathy, there is a notable gap in the literature regarding the specific investigation of neuronal ASIC expression in the brain. ⁶ We found only one study demonstrating marked neuronal ASIC1 expression reported only in the piglet striatum, where ASIC1 immunoreactivity was present in both striatonigral and striatopallidal neurons. ²⁴ Our present findings extend these findings showing that ASIC1 ion channels are also widespread in the neocortical areas with numerous ASIC1 expressing neurons in virtually all cortical layers, and all assessed cortical regions in accordance with other mammalian species.^39,40 However, we didn’t find any significant ASIC1 immunoreactivity in the cerebral vessels in accordance with human cerebrovascular cell type-specific single cell RNA sequencing analysis results showing that less than 2% of any cerebrovascular cell types including vascular endothelium, pericytes, microglia or astrocytes express ASIC1 (https://twc-stanford.shinyapps.io/human_bbb) from. ⁴¹

The role of ASIC1A in the cerebrovascular response to hypercapnia was established by the seminal paper of Faraci et al. ²¹ observing the responses of the pial arterioles in the cerebral cortex of adult mice. Using neuron-selective ASIC1A knockouts and local administration of PcTx1, they demonstrated that pial arteriolar responses to graded hypercapnia were virtually abolished in the genetically altered mice and were greatly reduced in response to 10% CO₂ inhalation in the PcTx1-treated wild-type animals. However, no method to study coinciding changes in CoBF was employed. ²¹ We are not aware of any previous study testing the connection between neuronal ASIC1A activation and hypercapnia-induced cerebrovascular response in the newborn brain. However, ASICs were indeed shown to be likely contributors to anoxic neuronal depolarization and consequent neuronal injury during hypoxic-ischemic stress in the piglet, as PcTx1 treatment was shown to provide significant neuroprotection of striatal neurons after hypoxic/ischemic injury. ²⁴ In that study, striatal neurons were protected by PcTx1 administered into the lateral ventricle 20 min before the onset of hypoxic-ischemic stress, and using FITC-PcTx1, the penetration of the toxin into the basal ganglia was demonstrated. ²⁴ In the present study, through the ports of the cranial window we continuously applied PcTx1 onto the cortical surface for 30 min, therefore, we can safely assume that the toxin likely penetrated all layers of the cortex by the onset of the graded hypercapnia stimulus in the present study.

In our present study, we found that the CoBF response to 5% CO₂ inhalation was unaltered by PcTx1, however, the additional increase in CoBF when switching to 10% CO₂ was significantly reduced. This is in sharp contrast to the findings in the ASIC1A knockout mice that virtually did not respond with pial arteriolar vasodilation to either 5% or 10% CO₂. ²¹ Also, PcTx1 in wild-type mice also nearly abolished the pial arteriolar response to 10% CO₂, and although the pial arteriolar responses to 5% CO₂ with PcTx1 were not reported, we can assume that they were similarly abolished (the response to the 10% CO₂ was less than half than the response to 5% in the absence of PcTx1). The direct comparison of the findings of our present study and that of Faraci et al. ²¹ is complicated by the fact that pial arteriolar diameter changes cannot be easily converted to CoBF alterations or vice versa. Clearly, hypercapnia elicits concentration dependent CoBF increases and vasodilation in both the pial and intraparenchymal arterioles.^2,42 The vasodilatory response can depend also on the size and the localization of the arterioles. Smaller arterioles tend to dilate more than the thicker ones ⁴² and the response is layer-specific. ⁵ In our study, the measured piglet pial arterioles had a much larger baseline diameters in a wider range compared to the ones observed in the mice. ²¹ That size difference may explain why the arteriolar response and the effect of PcTx1 on the diameter changes appeared to be also more variable. The vasodilation is accompanied with parenchymal CoBF elevation that could be semi-quantitatively assessed with LASCA. We found that inhibition of ASIC1A with PcTx1 did not affect the CoBF response to 5% CO₂. In a previous study, we determined the changes in the brain extracellular pH (pH_brain) of the piglet cortex in response to graded hypercapnia. ⁴³ We reported that hypercapnia induced with 5% or 10% CO₂ resulted in gradual drops of pH_brain over the course of 5–7 min from the baseline of 7.32 to 7.15 and 6.98, respectively. ⁴³ Taking into consideration that the ASIC1A were reported to be opened only by acidotic shifts below pH 7.0, ⁴⁴ the observed selective inhibitory effect of PcTx1 on additional CoBF increases to 10% CO₂ is highly conceivable.

Removal of the toxin by washing the cranial window with aCSF did not restore the CoBF response to hypercapnia that coincided with an exaggerated pial arteriolar dilatory response. Unfortunately, we cannot be certain whether PcTx1 was fully removed and/or the original sensitivity of ASIC1A channels were restored. Indeed, an in vitro study showed that long PcTx1 exposure resulted in the complete inhibition and long recovery of ASIC1A channel currents due to the possible slow unbinding of the toxin and changing the channel’s steady-state desensitization curve. ³² The enhanced pial arteriolar dilatory response after the toxin removal is unexplained but may perhaps be associated with a compensatory neuronal-vascular mechanism triggered by the ASIC1A channel desensitization mediated by the compensatory activation of NMDA receptors, increases in Ca²⁺ ion leading to nNOS mediated vasodilation.^32,45

In the present study, we determined not only the cerebrovascular but the cortical neuronal response to repeated graded hypercapnia and PcTX1 yielding some interesting observations. Importantly, we confirmed the complex LFP response to graded hypercapnia in this piglet model published previously, marked by pronounced reduction in the PSD of the local field potentials during exposure to 10% CO₂. ¹⁸ Furthermore, we showed that PcTx1 altered both individual and populational neuronal electrophysiological changes to graded hypercapnia. For instance, the modest PSD elevation in the θ band to 5% CO₂ was diminished, even after the toxin removal, also suggesting incomplete removal of PcTx1 and or restoration of ASIC1A excitability. Interestingly, despite general reduction of local field potential PSD, some neurons mainly in layer IV responded to 10% CO₂ with increased firing, and the number of these neurons even increased (being present in all cortical layers) in the presence of PcTx1, showing that general inhibition of neuronal activity cannot be responsible for the attenuated CoBF response to 10% CO₂.

The study was performed using only male piglets similarly to previous studies^8,9 on the mechanism of hypercapnia-induced cerebrovascular changes in this model that limits the interpretation of the results. The newborn piglet is an accepted translational model of neonatal hypoxic-ischemic encephalopathy, and a recent study found that male piglets are significantly more vulnerable to hypoxic-ischemic stress than females. ⁴⁶ As cerebrovascular reactivity to hypercapnia is known to be attenuated by hypoxic-ischemic stress^47,48 injuries, further studies are required to assess if differences in the neurovascular regulatory mechanisms can account for these differences.

In conclusion, neuronal ASIC1A channels that are abundantly expressed in the cerebral cortex of newborn pigs are involved in a neuronal vascular response responsible for the hypercapnia-induced CoBF increases in this species but their role does not appear to be as exclusive as suggested by the previous study in mice. Further studies are warranted to decipher the interactions of neuronal ASIC1A, NMDA, and nNOS in the regulation of the cerebral microcirculation of the neonate.