Background: Electrical vasoconstriction is a promising approach to control blood pressure or restrict bleeding in non-compressible wounds. We explore the neural and non-neural pathways of electrical vasoconstriction in-vivo. Methods: Charge-balanced, asymmetric pulses were delivered through a pair of metal disc electrodes. Vasoconstriction was assessed by measuring the diameter of rat saphenous vessels stimulated with low-voltage (20 V, 1 ms) and high- voltage (150 V, 10 μs) stimuli at 10 Hz for 5 min. Activation pathways were explored by topical application of a specific neural agonist (phenylephrine, alpha-1 receptor), a non-specific agonist (KCl) and neural inhibitors (phenoxybenzamine, 25 mg/ml; guanethidine, 1 mg/ml). Acute tissue damage was assessed with a membrane permeability (live-dead) fluorescent assay. The Joule heating in tissue was estimated using COMSOL Multiphysics modeling. Results: During stimulation, arteries constricted to 41 ± 8% and 37 ± 6% of their pre-stimulus diameter with low- and high-voltage stimuli, while veins constricted to 80 ± 18% and 40 ± 11%, respectively. In arteries, despite similar extent of constriction, the recovery time was very different: about 30 s for low-voltage and 10 min for high-voltage stimuli. Neural inhibitors significantly reduced low-voltage arterial constriction, but did not affect high-voltage arterial or venous constriction, indicating that high-voltage stimuli activate non-neural vasoconstriction pathways. Adrenergic pathways predominantly controlled low-voltage arterial but not venous constriction, which may involve a purinergic pathway. Viability staining confirmed that stimuli were below the electroporation threshold. Modeling indicates that heating of the blood vessels during stimulation (< 0.2 °C)istoo lowtocause vasoconstriction. Conclusions: We demonstrate that low-voltage stimuli induce reversible vasoconstriction through neural pathways, while high-voltage stimuli activate non-neural pathways, likely in addition to neural stimulation. Different stimuli providing precise control over the extent of arterial and venous constriction as well as relaxation rate could be used to control bleeding, perfusion or blood pressure. Keywords: Electrical stimulation, Electroceuticals, Vasoconstriction Background adenosine tetraphosphate  from endothelial cells, cir- For decades, electrical stimulation of cardiac striated culating hormones (i.e. angiotensin II) [9, 10], and dam- muscle has been successfully utilized in pacemakers and aged platelets . The dominant neural mechanisms defibrillators. Recently, electrical control of vascular smooth include a fast ionotropic (P2X receptor) pathway acti- muscle has been proposed to treat bleeding in vated by adenosine triphosphate (ATP) [12–15]; slower non-compressible wounds [1–3]. Understanding the metabotropic (alpha-1 and -2 adrenoreceptor) pathways vasoconstriction pathways activated by electrical stimuli will with the release of norepinephrine [15–21]; and the help create safe and effective devices for electrical control release of neuropeptide Y, which potentiates constriction of blood vessels. from norepinephrine and ATP [22–25]. Constriction of blood vessel involves both neural and Electrical stimulation of blood vessels has been used to non-neural pathways. Non-neural vasoconstriction study the neural pathways in constriction [15–18, 26–29] mechanisms include mechanical stretching (myogenic) and dilation [30, 31], including identification of norepin- [4, 5], release of endothelin-1 [6, 7] and uridine ephrine and ATP involvement by in-vitro stimulation of the rat saphenous artery . Direct electric current * Correspondence: email@example.com induces vasoconstriction and thrombosis [32–34]butalso Department of Bioengineering, University of Utah, 20 S. 2030 E., Salt Lake causes tissue damage. In-vitro studies have demonstrated City, UT 84112, USA both neural and non-neural electrically induced Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Brinton et al. Journal of NeuroEngineering and Rehabilitation (2018) 15:43 Page 2 of 10 vasoconstriction using a variety of vessel types (pulmonary, animals from each group—control (DMSO), phenoxyben- somatic and umbilical); however, they used direct or sinus- zamine and guanethidine—received about 200 μl of potas- oidal alternating current, which can damage tissue, and sium chloride (25 mg/ml, Sigma Aldrich; n =5) and, after could not directly compare the arterial and venous several minutes of washout, one drop of phenylephrine responses to the same stimulation because the vessels were HCl (25 mg/ml, Akorn; n = 5 for the Guanethidine block harvested from different locations [28, 29]. and n = 7 for the phenoxybenzamine block) as positive Based on our previous studies of vasoconstriction controls of vasoconstriction. Potassium chloride acts thresholds as a function of pulse duration, frequency, directly to depolarize nerves and smooth muscle while and amplitude [1, 2], we hypothesized that two distinct phenylephrine selectively activates the alpha-1 adrenergic stimuli (20 V, 1 ms and 150 V, 10 μs pulses at 10 Hz) receptors on smooth muscle cells. Because maximal could constrict the rat saphenous artery to a similar constriction occurred within several seconds, only the extent. Using these stimuli, we sought to compare the maximal constriction from phenylephrine and potassium extent and recovery time of arterial and venous constric- chloride are reported. tion in-vivo, describe the underlying pathways, and determine whether the stimulation required for vasocon- striction damages the rat saphenous vessels. Vessel stimulation and data collection Electrical stimulation and video monitoring of the vessels’ Methods width were performed as previously described . Briefly, Animals stainless steel disc electrodes, 1.6 mm diameter, were Male, Long Evans rats (Charles River), aged 50–60 days, placed 3.5 mm apart, with the saphenous vessels between with average weight of 309 g (range: 220-380 g) were used them. An anodic square pulse from a customized pulse in this study with approval by the Stanford Administrative generator was delivered through an 11 μF capacitor to the Panel on Laboratory Animal Care. Fourteen animals were electrodes to assure charge balance. Electrical parameters used to confirm the maximum constriction using the pro- were selected based on previous studies [1, 2], which dem- posed electrical parameters. In the vasoconstriction and onstrated that 150 V, 10 μs stimulation at 1 Hz induced a neural inhibition study 7 animals were used for control, 6 maximum constriction (about 30% of the original diam- for phenoxybenzamine and 5 for guanethidine blockade. eter), and 20 V, 1 ms stimulation at 1 Hz produced con- Additional five animals were used to assess vessel damage striction to 40–45% of the original diameter. Since with the live-dead assay. Before surgery, animals were anes- constriction also increased with pulse frequency, we hy- thetized using ketamine HCl (75 mg/kg) and xylazine pothesized that 20 V, 1 ms (referred to as low-voltage) and (5 mg/kg), with an additional half dose given every 45 min 150 V, 10 μs (referred to as high-voltage) stimulation, thereafter. pulsed at 10 Hz, would both reach the state of maximum For surgery, the animal was placed in the supine pos- constriction in the arteries. We first tested these stimula- ition and the rectal temperature was kept at 37 ± 1 °C. tion parameters in 14 animals (7 animals with both 20 V The saphenous artery and vein were exposed by removing and 150 V, 5 with only 150 V and 2 with only 20 V stimuli) the skin. Hartman’s Lactated Ringer solution (~ 37 °C) without pharmacological treatment (Fig. 1). When mul- dripped at about 1 Hz onto the surgical site during the tiple stimulations were delivered on the same animal, the surgery and stimulation. second occurred at least 15 min later and about 1 cm proximal to first stimulation. Stimulations lasted for Neural inhibitors 5 min, and the waveforms were monitored using an oscil- Neural inhibitors were applied topically once the skin was loscope (Tektronix, TDS 210). The low-voltage stimulus removed and vessels exposed. Phenoxybenzamine HCl delivered 12.5 mA (250 μJ/pulse) and the high-voltage (Sigma Aldrich) was dissolved in DMSO (25 mg/ml, stimulus 120 mA (180 μJ/pulse), as measured with a Sigma Aldrich). Guanethidine monosulfate (MyBioSource) 100 Ω series resistor. The inner diameter of blood vessels was first dissolved in de-ionized water (10 mg/ml) and was measured in ImageJ (NIH) from video captured with then diluted with DMSO to 1 mg/ml. Control animals a digital camera (Sentech Inc., TC202USB-A). (n = 7) received DMSO without inhibitors. The solutions Data were normalized by the vessel diameter prior to were applied liberally (~ 200 μl) to the exposed vessels and stimulation and presented as mean ± stdev. Statistical covered with a thin piece of plastic to prevent desiccation significance (p < 0.01) was determined using one-way (n = 6 for phenoxybenzamine and n = 5 for guanethidine). ANOVA, and, where appropriate, the post hoc Fresh solution was added (~ 200 μl) about every 5 min for Tukey-Kramer test to determine statistical significance a total of 25 min. The superficial fascia was removed to between study groups. To compare vessel constriction improve visualization of the vessel diameters prior to elec- between study groups, the vessel diameters were trical stimulation. Following electrical stimulation, some averaged over the 5-min stimulation period. Brinton et al. Journal of NeuroEngineering and Rehabilitation (2018) 15:43 Page 3 of 10 ab Fig. 1 Arterial and venous constriction in response to 10 Hz stimulation at 20 V, 1 ms and 150 V, 10 μs pulses. a Both stimuli constrict arteries to similar extent, but arteries dilate faster after low-voltage stimulation (*p < 0.01, F(3,34) = 23.24; one-way ANOVA with Tukey-Kramer multi- comparison test). b Veins constrict similarly to arteries at high voltage, but significantly less at low-voltage stimulation (p < 0.01, F(3,34) = 27.31; one-way ANOVA with Tukey-Kramer multi-comparison test on the average vessel diameter during stimulation); N = 7 for 20 V and N = 12 for 150 V stimuli Viability staining and vessel damage 0.7 cm. The tissue was assumed to have electrical and To determine whether the electrical stimulation damaged thermal properties of a muscle, and a 0.1 cm thick saline cell membranes, the saphenous vessels were surgically ex- layer covered the 0.6 cm thick muscle layer. Blood vessels posed, as described above, and the femoral artery of anes- in the muscle layer were modeled as 0.6 cm long cylinders thetized rats was cannulated with a micro-renethane of blood. Because constricted vessel diameters (especially catheter (Braintree Scientific, MRE-025) . The catheter the vein) depend on the electrical stimulus applied, we was advanced to the sapheneous branch, at which time modeled the vessel diameters and flow rates according to the animal was euthanized. Immediately, the distal end of our previous work (Table 1). The vessels passed be- the saphenous vessel was cut and the vessel flushed with tween two 0.16 cm diameter electrodes, separated by Ringers solution. A viability/cytotoxicity assay (BIOTUM, 0.35 cm, center-to-center. Electrodes were placed in direct 30002-T) was pumped through the vessel at a rate of contact with the muscle tissue, covered with saline. 0.02 ml/min (New Era Pump Systems, NE-300). While the To model the electric field in the tissue, we applied volt- staining solution flowed through the vessel, electrical age pulses and chose tissue conductivity so that the total pulses (20 V, 1 ms; 150 V, 10 μs; and 300 V, 10 μs) were injected current matched the current measured in-vivo applied at 10 Hz for 2 min to the exposed vessel in as- (12.5 mA and 120 mA for the 20 V and 150 V stimuli). cending order of voltage, moving the electrodes by 5 mm The in-vivo currents delivered through the electrodes for each voltage setting, along the saphenous vessels. were measured using a 100-Ohm resistor in series with Stimulation at 300 V, 10 μs was included as a positive con- the electrode and tissue. The current was measured at trol to validate the viability/cytotoxicity assay. After the pulse onset, before the capacitive interface had charged. last stimulation, the assay continued to flow through the The electrode surfaces were defined as equipotential, vessel for another 15 min. The vessel was then again and all other model boundaries were insulating. To flushed with Ringers solution, excised and mounted on a calculate the average Joule heating, the power density glass slide for imaging. The cytotoxicity component of the from one electrical pulse was multiplied by the duty assay (ethidium homodimer III) only enters the cells with cycle of the pulsed stimulus (0.01 for 1 ms pulses and damaged membranes to label the nuclei red, while the via- 0.0001 for 10 μs pulses). To account for the recharge bility component (calcein AM) crosses uncompromised Table 1 In-vivo vessel diameters and flow rates used in the cell membranes to label the entire viable cell green. The COMSOL modeling number of damaged cells (red) for each case were counted Applied Stimulus Artery Vein using ImageJ and normalized by the control. Constricted Diameter (μm) 20 V, 1 ms 200 430 Multiphysics modeling Based on the electric current measured in-vivo, we mod- 150 V, 10 μs 160 180 eled the electric field and tissue heating using COMSOL Flow Rate  (m/s) Multiphysics. The model assumed symmetry with respect 20 V, 1 ms 0.15 0.12 to the plane passing through the middle of the disk elec- 150 V, 10 μs 0.09 0.07 trodes to reduce the modeling volume to 1.5 × 0.5 × Brinton et al. Journal of NeuroEngineering and Rehabilitation (2018) 15:43 Page 4 of 10 phase of the pulse, we took the conservative approach walls of the control (no stimulation), as well as in the and doubled this average power density. In muscle, a 20 V and 150 V samples. However, with the 300 V stim- perfusion term was included according the bioheat equa- uli (10 μs pulses at 10 Hz), the number of damaged cells tion . Blood flow rates through the vessels were esti- increased 4.7-fold (p < 0.05) and the arterial wall dilated mated using the in-vivo measured diameters and beyond its non-stimulated diameter. published bleeding rates of electrically constricted saphe- nous vessels in rats of similar weight and electrical Tissue heating by electrical stimulation parameters (Table 1). Blood flowed into the model at To assess the extent of tissue heating during stimulation, 37 °C from the rear in both vessels, a simplification justi- we modeled the system using COMSOL Multiphysics. fied by the minimal blood heating due to high flow rates. At steady-state, about 25 s of stimulation, the The boundaries of the modeled volume were set to 37 ° temperature on electrode surface increases by about C to reflect body temperature and the drip of the warm 0.96 °C with 20 V and by 0.61 °C with 150 V stimuli saline onto the vessels. The muscle and saline boundar- (Fig. 3). Near blood vessels, the steady-state temperature ies in the plane of symmetry were insulating. The elec- rise was only 0.2 °C and 0.15 °C from the 20 V and trical and thermal material properties are detailed in 150 V stimuli, respectively. As a worst-case scenario, we Table 2. The tissue temperature reached a steady state modeled tissue heating without blood flow in the vessels, within about 25 s of stimulation. For comparison, the which yielded temperature rise of 0.6 °C and 0.35 °C at tissue heating without blood flow through the vessels the vessel walls. Both temperatures are well within the was also calculated. range of physiological variations, and less than the temperature change from each drip of warmed saline on Results the vessel surface. Electrical stimulation of blood vessels Upon electrical stimulation, arteries constricted to 41 ± 8% Neural pathways and 37 ± 6% of the initial vessel diameter with 20 V, 1 ms To determine whether neural and non-neural pathways and 150 V, 10 μspulsesrepeatedat10 Hz(Fig. 1(a)). While were involved in electrically-induced vasoconstriction, we arteries constricted to a similar extent with both stimuli, applied a selective neural agonist (phenylephrine), neural vessels treated with low-voltage dilated back to 90% of the inhibitors (phenoxybenzamine and guanethidine), and a initial diameter within 30 s, while vessels treated with non-specific depolarizing agent (potassium chloride). high-voltage stimulation recovered after 10 min. Phenylephrine is a synthetic analog of norepinephrine that With high-voltage stimulation, veins constricted to a binds and activates the alpha-1 adrenergic receptor. similar extent as arteries (40 ± 11% of the initial vessel Phenoxybenzamine (PBZ) prevents binding to the alpha-1 diameter), but to only 80 ± 18% of the initial diameter and alpha-2 adrenergic receptors on smooth muscle cells. with low-voltage stimuli (Fig. 1(b), p < 0.01). Veins also Guanethidine prevents the release of adrenergic (norepin- recovered slower than arteries—10 min with low-voltage ephrine) and purinergic (adenosine triphosphate) neuro- and 15 min with high-voltage stimuli. transmitters from sympathetic nerves. Potassium chloride (KCl) depolarizes both nerves and smooth muscle cells Assessment of tissue damage by electrical stimulation directly. Since maximal constriction occurred immediately To evaluate tissue damage by electrical stimulation following application, we plot these levels of constrictions (electroporation), we applied a fluorescent cell viability as horizontal lines in Fig. 4, for comparison. assay to stimulated blood vessels: cells with perme- abilized membranes fluoresce in red, while intact cells Neural inhibition: Arteries are stained with green. As shown in Fig. 2, a limited As expected, phenylephrine induced a strong arterial con- number of damaged cells (red) can be seen in the vessel striction to 45 ± 16% of the initial vessel diameter (Fig. 4, green dash). Pretreatment with PBZ completely blocked arterial constriction with phenylephrine—the vessel con- Table 2 Electrical and thermal parameters for COMSOL modeling stricted to only 96 ± 5% of the initial diameter (p <0.01; Fig. 4(a) and Fig. 4(c), green solid). Potassium Chloride Muscle Blood Saline (KCl) produced the largest arterial constriction, down to Electrical Conductivity (S/m) 0.376 (20 V) 0.76 [52, 53] 1.0 0.475(150 V) 25 ± 2% of the initial diameter (Fig. 4, black dash). PBZ slightly reduced the constrictive effect of KCl, presumably Thermal Conductivity (W/m∙K) 0.52  0.5  0.59 by preventing the endogenous norepinephrine, released Specific Heat Capacity (J/kg∙K) 3550  3840  4173  from depolarized nerves, from binding to the adrenergic Density (kg/m ) 1041 1055 1000 receptors (alpha-1 and alpha-2) on the smooth muscle − 1 Perfusion Parameter (s ) 0.00067 NA NA cells; however, KCl can also act on smooth muscle directly Brinton et al. Journal of NeuroEngineering and Rehabilitation (2018) 15:43 Page 5 of 10 ab cd Fig. 2 Cell viability assay of the stimulated artery. Green color indicates intact cells, while red shows cells with compromised cell membrane. a Control with no stimulation (N = 5); b 20 V, 1 ms pulses (N = 5); c 150 V, 10 μs pulses (N = 5); or d 300 V, 10 μs pulses (N = 4). All treatments were applied at 10 Hz for 2 min. The number of compromised cells was normalized by the control. Vessels stimulated by 20 V and 150 V pulses exhibited damage similar to the control (1.1 ± 0.8, 1.2 ± 0.3, and 1.0 ± 0.6 a.u., respectively), while the 300 V stimulation damaged 4.7-fold more cells (4.7 ± 1.6 a.u.; p < 0.05 and F(3,15) = 18.15, using one-way ANOVA and Tukey-Kramer multi-comparison tests) ab cd Fig. 3 Temperature rise from a 20 V, 1 ms and c 150 V, 10 μs stimulations, pulsed at 10 Hz. The temperature rise along the line passing through the center of blood vessels for b 20 V and d 150 V stimulation Brinton et al. Journal of NeuroEngineering and Rehabilitation (2018) 15:43 Page 6 of 10 ab cd Fig. 4 Neural inhibition of arterial vasoconstriction. a Phenoxybenzamine (PBZ) completely blocked constriction induced by phenylephrine (PE, p < 0.01), slightly inhibited constriction induced by potassium chloride (KCl, p = 0.02), and significantly decreased constriction induced by 20 V stimulation (*p < 0.01, F(11,58) = 98.75; one-way ANOVA and Tukey-Kramer multi-comparison test). b Guanethidine had no effect on vasoconstriction induced by PE or by KCl, but it blocked the 20 V-induced constriction (*p < 0.01). c PBZ did not inhibit constriction induced by 150 V stimulation, demonstrating that vasoconstriction by high-voltage bypassed the adrenergic neural pathway. d On average, guanethidine failed to block constriction induced by 150 V stimuli. However, the onset of constriction was slowed down (*p < 0.01; one-way ANOVA, F(1,9) = 25.14). Unless specified, significance was determined using one-way ANOVA and Tukey-Kramer multi-comparison test, F(11,58) = 98.75; N = 5 for guanethidine, N =6 for PBZ and N = 7 control. Horizontal lines indicate the maximum constriction achieved with phenylephrine (green) and potassium chloride (black) so the arterial diameter increased only slightly from With high-voltage stimulation, PBZ and guanethidine 25 ± 2% to 37 ± 4% of the initial (p < 0.01; Fig. 4(a)and failed to block arterial vasoconstriction: the vessel diame- Fig. 4(c), black). As expected, Guanethidine failed to ters were 30 ± 3% of the initial without PBZ or guanethi- inhibit constriction induced by phenylephrine because dine, 30 ± 6% with PBZ, and 37 ± 2% with guanethidine phenylephrine acts downstream of the guanethidine (Fig. 4(c) and Fig. 4(d), red). However, in the presence of blockade: 45 ± 16% (without guanethidine) compared with guanethidine, arterial constriction occurred slower: at 30 s 40 ± 3% (with guanethidine) of the initial arterial diameter the artery constricted to 28 ± 4% of the initial diameter in (Fig. 4(b)and Fig. 4(d), green). Guanethidine also had a lim- the control but just 51 ± 11% of the initial diameter with ited effect on KCl, which can act directly on smooth muscle guanethidine (p < 0.01; Fig. 4(d), red). cells and constricted the vessel to 25 ± 2% of the initial ves- sel diameter without guanethidine, compared with 30 ± 4% Neural inhibition: Veins with guanethidine (Fig. 4(b)and Fig. 4(d), black). The neural agonists and antagonists affected veins quite Pretreatment with PBZ reduced arterial constriction differently. Phenylephrine (PE) failed to elicit a venous with low-voltage (20 V) stimuli and increased the vessel contraction and was not affected by the neural inhibitors: diameter from 38 ± 4% (without PBZ) to 79 ± 7% (with the vein diameters were 97 ± 13% (PE alone), 102 ± 7% PBZ) of the original (p <0.01; Fig. 4(a), red). Pretreatment (with PBZ) and 92 ± 10% (with guanethidine). PBZ re- with guanethidine completely eliminated arterial constric- duced venous constriction with KCl and increased the tion with low-voltage (20 V) stimulation—the stimulated vessel diameter from 51 ± 21% of the initial to 95 ± 6% vessel diameter increased from 38 ± 4% (without guanethi- with PBZ (p = 0.02). The increase to 74 ± 19% with dine) to 98 ± 2% (with guanethidine) of the pre-stimulus guanethidine was not significant (Fig. 5,black). value (p <0.01; Fig. 4(b), red). Guanethidine pretreatment PBZ did not affect the venous response to low-voltage also revealed a post-stimulus dilation up to 127% of the stimulation (78 ± 9% of the pre-stimulus diameter com- initial diameter, which trends back to normal over 15 min. pared with 73 ± 15% with PBZ), while guanethidine Brinton et al. Journal of NeuroEngineering and Rehabilitation (2018) 15:43 Page 7 of 10 ab cd Fig. 5 Neural inhibition of venous vasoconstriction. a PBZ did not block venous constriction induced by 20 V stimulation, but it reduced constriction induced by KCl (p = 0.02). b Although guanethidine reduced constriction induced by 20 V stimulation and by KCl, the differences were not significant. c PBZ did not affect vasoconstriction induced by 150 V stimulation. d Guanethidine did not affect constriction induced by 150 V stimulation. Horizontal lines indicate the maximum constriction achieved with potassium chloride (black). Statistical comparisons were performed using the one-way ANOVA and the Tukey-Kramer multi-comparison test; N = 5 for guanethidine, N = 6 for PBZ and N = 7 for control reduced constriction, but not significantly (94 ± 10% of the durations shorter than the typical cell polarization time pre-stimulus diameter) (Fig. 5(a) and Fig. 5(b), blue). (~ 50 ns) can selectively polarize intracellular organelles Neither PBZ nor guanethidine affected venous constriction . While damage of intracellular organelles would not by high-voltage stimulation: vein diameters were 36 ± 4% be detected with our viability assay, it is unlikely that of the pre-stimulus diameter (control), compared with our stimulus induced any direct action on them because 37 ± 4% (PBZ) and 36 ± 5% (guanethidine) (Fig. 5(c)and electric field in our case is about 2 orders of magnitude Fig. 5(d), blue). lower than that required for activation of intracellular organelles. Extremely high electric fields can also acti- Discussion vate platelets in the blood, which may constrict vessels Electrical stimulation of blood vessels by releasing thromboxane or serotonin [11, 38, 39], but We found that both, high- and low-voltage stimuli con- this pathway is also unlikely due to the substantially strict saphenous arteries to a similar extent in-vivo, but lower electric fields in our study. low-voltage engages a neural pathway that recovers With 150 V, 0.01 ms stimuli, charge density at the quickly (within 30 s), while high-voltage activates a electrode surface (60μC/cm ) is close to the capacitive non-neural pathway that recovers slowly (over several coupling limit for stainless steel-electrolyte interface minutes). We also show that high-voltage stimulation (40-50μC/cm ), and may be delivered without electroly- constricts veins as much as arteries, but low-voltage sis due to surface roughness. For the 20 V, 1 ms stimu- constricts only half that amount. These observations lus, charge density (625μC/cm ) exceeds the capacitive suggest that different vasoconstriction pathways could coupling limit, so the current was sustained via elec- be activated by electrical stimulation. trolysis of water . However, even with electrolysis, it is unlikely that gas byproducts or changes in pH affected Electrical stimulation below damage threshold vasoconstriction since the electrodes were located sev- Strong electric field can permeabilize and damage cell eral millimeters away from the vessels and warm saline membranes; however, our cellular viability assay showed was continuously washing the tissue surface. To avoid no damage to arteries with the 20 and 150 V stimuli. hydrolysis in clinical applications, electrodes should have Extremely high electric field (tens of kV/cm) with pulse sufficiently high capacitance, such as sputtered iridium Brinton et al. Journal of NeuroEngineering and Rehabilitation (2018) 15:43 Page 8 of 10 oxide films (SIROF), which can safely supply charge blockage of vasoconstriction with guanethidine) which densities exceeding 1mC/cm . confirmed ex-vivo observations . PBZ completely eliminated venous constriction by KCl, while it was only slightly reduced in case of arterial con- Heating by electrical stimulation striction by KCl. This implies that KCl induces venous For electric field modeling, we selected the muscle conduct- constriction by depolarizing neurons that release norepin- ivity, so that the total current matched the in-vivo mea- ephrine. Because phenylephrine, a pure alpha-1 agonist, sured current (12.5 mA or 120 mA for the 20 or 150 V did not affect the vein, we conclude that saphenous vein electrodes). This approach resulted in a slightly lower value constriction occurs primarily through the alpha-2 recep- of muscle conductivity for the 20 V stimulus (likely due to tors, which are activated by norepinephrine, blocked by gas formation at the electrode-electrolyte interface). Our PBZ and unaffected by phenylephrine. The alpha-2 recep- thermal modeling demonstrated a temperature rise below tor pathway was also shown to be the dominant venous 1°Conelectrodes, andonly0.15–0.2°Conthe vessel walls, constriction pathway in dogs [19, 20]. Interestingly, the even without considering cooling from convection at the adrenergic pathway (alpha-1 and -2 receptors) does not exposed saline surface. Such a minimal heating is very un- appear to be involved in low-voltage venous constriction likely to induce vasoconstriction since temperature pulsa- because pretreatment with PBZ failed to block constric- tion by a few degrees from a drip of warm saline (37 °C) tion. Low-voltage venous constriction may involve activa- did not affect the vessel diameter. The vessel heating is tion of a purinergic pathway because veins treated with similar to our previous reported values (about 2.5 °C with guanethidine constricted less than without purinergic 150 V, 100 μspulsesat10 Hz and 0.02 °C with 80 V, 1 μs blockage (Fig. 5(b)). pulses at 10 Hz) ; however, our current model predicts Low-voltage, neural stimulation primarily affects arter- even less heating with 150 V pulses because of about ial constriction and flow, which could be useful to 10-fold less charge per pulse and luminal blood flow based control hemorrhage , blood perfusion or blood pres- on in-vivo measurements. While variations in blood flow sure in a localized tissue or organ. The neural pathway affect the modeling results, even without blood flow, the provides rapid constriction and dilation and can safely vessels will heat no more than 0.6 and 0.35 °C for 20 V and constrict vessels for hours . However, chronic stimula- 150 V stimuli—again less temperature variation than that tion will require electrode materials capable of safely produced by the dripping 37 °C saline. injecting 625μC/cm , such as SIROF or TiN [40, 41]. Arterial dilation following low-voltage stimulation was Neural pathways observed most clearly in guanethidine treated vessels The thermal modeling and cell viability assay suggest (Fig. 4(b)), and it may be mediated by release of nitric that vasoconstriction was not induced by electroporation oxide or prostaglandins [42, 43]. Because the dilation or vessel heating. To understand the mechanisms of presented only when the neurotransmitters were electrical vasoconstriction we applied the pharmaco- blocked, the dilatory effect appears to be overpowered logical inhibitors PBZ and guanethidine. PBZ partially under normal stimulation conditions (no pharmaco- blocks neuro-mediated constriction by preventing the logical blockade). Further studies could determine neurotransmitter (norepinephrine) from binding to whether this effect could be exploited to increase blood alpha-1 and alpha-2 receptors on the smooth muscle flow in tissue with poor circulation. cells , while guanethidine provides a complete neural block by preventing the release of adrenergic (norepin- Neural inhibition during high-voltage stimulation ephrine) and purinergic (adenosine triphosphate) neuro- In-vivo, high-voltage vasoconstriction was not dependent transmitters from sympathetic nerves . on a neural pathway, since it was not affected by neuro- transmitter blockers and confirms previous in-vitro studies Neural inhibition during low-voltage stimulation showing both arterial and venous constriction in the pres- In-vivo, low-voltage constriction in arteries was ence of neural inhibitors [28, 29]. Direct depolarization of neuro-mediated, with about 65% of the effect due to the smooth muscle with high-voltage stimuli is unlikely be- adrenergic pathway and additional 30–35% from the puri- cause high-voltage constriction persists for several minutes nergic pathway, as evidenced by the partial and complete after stimulation, unlike KCl-induced constriction which inhibition with PBZ and guanethidine, respectively. Neural directly depolarizes smooth muscle and reverses within a vasoconstriction through adrenergic (dominant) and puri- minute of rinsing the solution. Furthermore, it has been nergic pathways was also observed ex-vivo in rat saphe- shown that contractility of smooth muscle decreased rap- nous arteries using similar electrical parameters . idly below 165μC/cm per pulse at 20 Hz . Our Low-voltage stimulation did not depolarize the arterial high-voltage stimulation generates 8-fold less charge dens- smooth muscle directly in-vivo (evidenced by complete ity per pulse (20μC/cm ) at the arterial wall with half the Brinton et al. Journal of NeuroEngineering and Rehabilitation (2018) 15:43 Page 9 of 10 pulse frequency (10 Hz), further indicating that a direct ef- nerves could be reduced by using sensors to identify and fect on smooth muscle is unlikely in our case. stimulate only near the source of hemorrhage, by using High-voltage electrical vasoconstriction may result neuromuscular blocking agents available during general from release of endothelin-1 by endothelial cells in the anesthesia or by intermixing the vasoconstriction stimuli lumen of arteries and veins: endothelin-1 constricts ves- with high pulse frequency stimuli, capable of blocking the sels to a similar extent as KCl, and does not readily pain or completely exhausting the neuromuscular wash-out (vessels remain constricted for more than junction. 10 min) [6, 7, 45, 46]. Endothelial cells under mechanical stress can also release uridine adensosine tetraphosphate Conclusions and induce potent vasoconstriction . Since vasocon- Pulsed electrical stimulation provides a reversible and striction is localized between the electrodes, circulating non-damaging approach to blood vessel control in-vivo. agents (such as angiotensin) are unlikely to play a role Low-voltage stimuli engage neural vasoconstriction because they would diffuse downstream rather than con- pathways, while high-voltage also activates non-neural strict the vessel only locally. pathways to induce maximum arterial constriction. The For some applications, high-voltage, non-neural vaso- low- and high-voltage stimuli provide different extent of constriction has the advantage of constricting veins constriction and rates of dilation, which could be useful nearly as much as arteries. This could help control trau- in a variety of applications for control of bleeding, perfu- matic bleeding in highly perfused tissue, where the sion, or blood pressure. major arterial blood supply may be difficult to locate or Abbreviations reach, or in sacral and pelvic cavities where venous KCl: Potassium chloride; PBZ: Phenoxybenzamine hemorrhage can be significant [47–49]. Since high-voltage stimulation uses 40% less energy Funding This work was funded by the U.S. Department of Defense, Air Force Office of per pulse, achieves maximum constriction with 10-fold Scientific Research (FA9550–14–1-0074). lower pulse frequency , and could be applied intermit- tently because constriction lasts several minutes, it could Availability of data and materials The datasets analyzed during the current study are available from the enable smaller, more power efficient devices for long corresponding author on request. lasting vessel control. At 1 Hz, high-voltage delivers 14-fold less power than the low-voltage stimulation. Authors’ contributions All authors conceived experiments and helped to draft the manuscript. MB performed the experiments and analyzed the data. All authors read and Limitations approve the final manuscript One limitation of this study is that we have not shown safety for clinically relevant durations of stimulation (i.e. Competing interest The authors declare that they have no competing interests. greater than 30 min). However, histological examination of the rat saphenous vessels showed no vessel damage Ethics approval one week after a 60-min-long stimulation with identical This study was approved by the Stanford Administrative Panel on Laboratory electrodes at low voltage (20 V, 1 ms pulses at 10 Hz) Animal Care. . In addition, a previous study demonstrated that the threshold of cellular damage by electroporation does not Publisher’sNote Springer Nature remains neutral with regard to jurisdictional claims in decrease beyond about 50 pulses, suggesting that longer published maps and institutional affiliations. stimulations should also be safe . The DMSO used in the inhibitor experiments extended Author details Department of Bioengineering, University of Utah, 20 S. 2030 E., Salt Lake the arterial recovery time after high-voltage constriction City, UT 84112, USA. Faculty of Life Sciences, Bar Ilan University, 5290002 (comparing Fig. 1(a) and Fig. 4(c)). However, it did not Ramat-Gan, Israel. Department of Ophthalmology, Stanford University, 2452 affect the extent of constriction, so comparisons between Watson Court Palo Alto, Stanford, CA 94303, USA. Hansen Experimental Physics Laboratory, Stanford University, 452 Lomita Mall, Stanford, CA 94305, neural inhibitors and their controls are accurate. Even USA. without DMSO (Fig. 1(b)), the vein did not fully dilate 15 min after high-voltage stimulation, perhaps due to Received: 12 February 2018 Accepted: 22 May 2018 lower blood pressure compared with the artery. Electrical stimulation capable of inducing vasoconstric- References tion also activates the nearby muscles and sensory nerves. 1. Mandel Y, et al. “Vasoconstriction by electrical stimulation: new approach to Due to depletion of acetylcholine at the neuromuscular control of non-compressible hemorrhage,” (in eng). Sci Rep. 2013;3:2111. 2. Brinton MR, Mandel Y, Dalal R, Palanker D. “Miniature electrical stimulator for junction, muscle contraction with each stimulus pulse hemorrhage control,” (in eng). IEEE Trans Biomed Eng. 2014;61(6):1765–71. decreased over time, and was almost gone after about a 3. Mandel Y, et al. Hemorrhage control of liver injury by short electrical pulses. minute. Unwanted activation of the muscle and sensory PLoS One. 2013;8(1):e49852. Brinton et al. Journal of NeuroEngineering and Rehabilitation (2018) 15:43 Page 10 of 10 4. Davis MJ, Hill MA. “Signaling mechanisms underlying the vascular myogenic 30. Hughes J, Vane JR. Relaxations of isolated portal vein of rabbit induced by response,” (in eng). Physiol Rev. 1999;79(2):387–423. nicotine and electrical stimulation. Br J Pharmacol. 1970;39(3):476–89. 5. Schubert R, Lidington D, Bolz SS. “The emerging role of Ca2+ sensitivity 31. Torok J, Zemancikova A. Relaxation of rat main pulmonary artery to electrical regulation in promoting myogenic vasoconstriction,” (in eng). Cardiovasc stimulation: role of nitric oxide. Act Nerv Super Rediviva. 2010;52(2):151–6. Res. 2008;77(1):8–18. 32. Guarini S. A highly reproducible model of arterial thrombosis in rats. J 6. Furchgott RF, Vanhoutte PM. “Endothelium-derived relaxing and contracting Pharmacol Toxicol Methods. 1996;35(2):101–5. factors,” (in eng). FASEB J. 1989;3(9):2007–18. 33. Hladovec J. “Experimental arterial thrombosis in rats with continuous registration,” (in eng). Thromb Diath Haemorrh. Oct 1971;26(2):407–10. 7. Tykocki NR, Watts SW. “The interdependence of endothelin-1 and calcium: a 34. Bourgain RH, F S. A continuous registration method in experimental arterial review,” (in eng). Clin Sci (Lond). 2010;119(9):361–72. thrombosis in the rat. Thromb Res. 1974;4:599–607. 8. Jankowski V, et al. “Uridine adenosine tetraphosphate: a novel endothelium- 35. Jespersen B, Knupp L, Northcott CA. “Femoral arterial and venous derived vasoconstrictive factor,” (in eng). Nat Med. 2005;11(2):223–7. catheterization for blood sampling, drug administration and conscious 9. Majumder K, Wu J. “Molecular targets of antihypertensive peptides: blood pressure and heart rate measurements,” (in eng). Vis Exp. 2012;(59). understanding the mechanisms of action based on the pathophysiology of doi: https://doi.org/10.3791/3496. hypertension,” (in eng). Int J Mol Sci. 2014;16(1):256–83. 36. Pennes HH. “Analysis of tissue and arterial blood temperatures in the resting 10. Zhou Y, Chen Y, Dirksen WP, Morris M, Periasamy M. “AT1b receptor human forearm,” (in eng). J Appl Physiol. 1948;1(2):93–122. predominantly mediates contractions in major mouse blood vessels,” (in 37. Vernier PT, Sun Y, Marcu L, Salemi S, Craft CM, Gundersen MA. “Calcium eng). Circ Res. 2003;93(11):1089–94. bursts induced by nanosecond electric pulses,” (in eng). Biochem Biophys 11. Golino P, et al. “Local platelet activation causes vasoconstriction of large Res Commun. 2003;310(2):286–95. epicardial canine coronary arteries in vivo. Thromboxane A2 and serotonin 38. Torres AS, et al. “Platelet activation using electric pulse stimulation: growth are possible mediators,” (in eng). Circulation. 1989;79(1):154–66. factor profile and clinical implications,” (in eng). J Trauma Acute Care Surg. 12. Kur J, Newman EA. “Purinergic control of vascular tone in the retina,” (in 2014;77(3 Suppl 2):S94–S100. eng). J Physiol. 2014;592(3):491–504. 39. Frelinger AL, et al. “Modification of pulsed electric field conditions results in 13. Tarasova O, Sjöblom-Widfeldt N, Nilsson H. “Transmitter characteristics of distinct activation profiles of platelet-rich plasma,” (in eng). PLoS One. 2016; cutaneous, renal and skeletal muscle small arteries in the rat,” (in eng). Acta 11(8):e0160933. Physiol Scand. 2003;177(2):157–66. 40. Merrill DR, Marom B, GR JJ. Electrical stimulation of excitable tissue: design 14. Burnstock G, Ralevic V. “Purinergic signaling and blood vessels in health and of efficacious and safe protocols. J Neurosci Methods. 2005;141(2):171–98. disease,” (in eng). Pharmacol Rev. 2014;66(1):102–92. 41. Cogan SF. Neural stimulation and recording electrodes. Annu Rev Biomed 15. Burnstock G, Warland JJ. “A pharmacological study of the rabbit saphenous Eng. 2008;10:275–309. artery in vitro: a vessel with a large purinergic contractile response to 42. Zhang W, Edvinsson L, Lee TJ. “Mechanism of nicotine-induced relaxation in sympathetic nerve stimulation,” (in eng). Br J Pharmacol. 1987;90(1):111–20. the porcine basilar artery,” (in eng). J Pharmacol Exp Ther. 1998;284(2):790–7. 16. Sjöblom-Widfeldt N, Nilsson H. “Sympathetic transmission in small 43. Tartas M, et al. Cathodal current-induced vasodilation to single application mesenteric arteries from the rat: influence of impulse pattern,” (in eng). Acta and the amplified response to repeated application in humans rely on Physiol Scand. 1990;138(4):523–8. aspirin-sensitive mechanisms. J Appl Physiol (1985). 2005;99(4):1538–44. 17. Ferrell WR, Khoshbaten A. Responses of blood-vessels in the rabbit knee to 44. van Mastrigt R, Glerum JJ. “Electrical stimulation of smooth muscle strips electrical-stimulation of the joint capsule. J Physiol. 1990;423:569–78. from the urinary bladder of the pig,” (in eng). J Biomed Eng. 1985;7(1):2–8. 18. Khoshbaten A, Ferrell WR. Nerve-mediated responses of blood-vessels in the 45. Burnstock G. “Endothelium-derived vasoconstriction by purines and rabbit knee-joint. J Vasc Res. 1993;30(2):102–7. pyrimidines,” (in eng). Circ Res. 2008;103(10):1056–7. 19. De Mey J, Vanhoutte PM. “Uneven distribution of postjunctional alpha 1- 46. Yanagisawa M, et al. “A novel potent vasoconstrictor peptide produced by and alpha 2-like adrenoceptors in canine arterial and venous smooth vascular endothelial cells,” (in eng). Nature. 1988;332(6163):411–5. muscle,” (in eng). Circ Res. 1981;48(6 Pt 1):875–84. 47. Lallemand MS, et al. “Resuscitative endovascular balloon occlusion of the 20. MacLennan SJ, Luong LA, Jasper JR, To ZP, Eglen RM. “Characterization of alpha aorta for major abdominal venous injury in a porcine hemorrhagic shock 2-adrenoceptors mediating contraction of dog saphenous vein: identity with model,” (in eng). J Trauma Acute Care Surg. 2017;83(2):230–6. the human alpha 2A subtype,” (in eng). Br J Pharmacol. 1997;121(8):1721–9. 48. Weber DG, Bendinelli C, Balogh ZJ. “Damage control surgery for abdominal 21. Dunn WR, McGrath JC, Wilson VG. “Postjunctional alpha-adrenoceptors in emergencies,” (in eng). Br J Surg. 2014;101(1):e109–18. the rabbit isolated distal saphenous artery: indirect sensitivity to prazosin of 49. Costantini TW, et al. “Current management of hemorrhage from severe responses to noradrenaline mediated via postjunctional alpha 2- pelvic fractures: results of an American Association for the Surgery of adrenoceptors,” (in eng). Br J Pharmacol. 1991;103(2):1484–92. Trauma multi-institutional trial,” (in eng). J Trauma Acute Care Surg. 2016; 22. Donoso MV, Steiner M, Huidobro-Toro JP. “BIBP 3226, suramin and prazosin 80(5):717–23; discussion 723–5. identify neuropeptide Y, adenosine 5′-triphosphate and noradrenaline as 50. Butterwick A, Vankov A, Huie P, Freyvert Y, Palanker D. Tissue damage by sympathetic cotransmitters in the rat arterial mesenteric bed,” (in eng). J pulsed electrical stimulation. IEEE Trans Biomed Eng. 2007;54(12):2261–7. Pharmacol Exp Ther. 1997;282(2):691–8. 51. Bronzino JD. The biomedical engineering handbook, 2nd ed. (electrical 23. Cortés V, et al. “Synergism between neuropeptide Y and norepinephrine engineering handbook series). Boca Raton, FL: CRC Press; 2000. p. 2. volumes highlights sympathetic cotransmission: studies in rat arterial mesenteric bed 52. Duck F. Physical Properties of Tissues: A Comprehensive Reference Book. with neuropeptide Y, analogs, and BIBP 3226,” (in eng). J Pharmacol Exp London: Academic Press; 1990. Ther. 1999;289(3):1313–22. 53. Duck FA. Physical properties of tissue : a comprehensive reference book. 24. Cheung DW. “Neuropeptide Y Potentiates specifically the purinergic London: Academic Press; 1990. component of the neural responses in the Guinea pig saphenous artery,” (in 54. K. Giering, I. Lamprecht, and O. Minet, “Specific heat capacities of human eng). Circ Res. 1991;68(5):1401–7. and animal tissues,” Proc SPIE 2624, Laser-Tissue Interaction and Tissue 25. Fabi F, Argiolas L, Ruvolo G, del Basso P. “Neuropeptide Y-Induced Optics, vol. 188, 1996. potentiation of noradrenergic vasoconstriction in the human saphenous 55. Fegler G. “Measurement of cardiac output in anaesthetized animals by vein: involvement of endothelium generated thromboxane,” (in eng). Br J a thermodilution method,” (in eng). Q J Exp Physiol Cogn Med Sci. Pharmacol. 1998;124(1):101–10. 1954;39(3):153–64. 26. Paterson G. “The response to transmural stimulation of isolated arterial strips 56. Christensen DA. Ultrasonic bioinstrumentation. New York: Wiley; 1988. pp. and its modification by drugs,” (in eng). J Pharm Pharmacol. 1965;17:341–9. xiii, 235 pages 27. Yates CM, Gillis CN. “The response of rabbit vascular tissue to electrical and drug stimulation,” (in eng). J Pharmacol Exp Ther. 1963;140:52–9. 28. Furchgott RF. “The pharmacology of vascular smooth muscle,” (in eng). Pharmacol Rev. 1955;7(2):183–265. 29. Somlyo AV, Somlyo AP. “Electromechanical and pharmacomechanical coupling in vascular smooth muscle,” (in eng). J Pharmacol Exp Ther. 1968; 159(1):129–45.
Journal of NeuroEngineering and Rehabilitation
– Springer Journals
Published: May 29, 2018