Systemic Hypoxia and Interventions on Nitric Oxide Metabolism Affect Retinal Vasomotion in Healthy Subjects Authors: Line Petersen, Toke Bek Published: 2024 ABSTRACT PurposePrevious in vitro studies on porcine retinal arterioles have shown that the frequency and amplitude of retinal vasomotion can be affected by hypoxia and nitric oxide (NO). However, it is unknown whether these effects can be reproduced in humans in vivo.MethodsVideo recordings of retinal arterioles from 40 healthy subjects were studied before and during breathing of a hypoxic gas mixture consisting of 12.5% oxygen and 87.5% nitrogen. The experiments were repeated before and after the addition of either an NO donor or continuous infusion of an NO synthesis inhibitor. The harmonic content of spontaneous diameter changes in retinal arterioles was extracted by Fourier analysis, and the effects of the interventions were studied on the amplitudes of diameter oscillations in the frequency bands 1: 0.5/min–3.5/min, 2: 4/min–7.5/min, 3: 8/min–17.5/min, and 4: 18/min–30/min.ResultsHypoxia significantly increased the amplitude of spontaneous diameter oscillations within the low (1) and high (3–4) frequency bands (P < 0.006 for all comparisons), and the effect in frequency band 1 was eliminated by the NO donor. In frequency band 2, hypoxia had no significant effect on the amplitude of diameter oscillations, but the amplitudes were significantly reduced by the NO donor (P < 0.01) and significantly increased by the NO synthesis inhibitor (P = 0.03).ConclusionsHypoxia and interventions on NO metabolism can affect spontaneous diameter oscillations in retinal arterioles. Disturbances in vasomotion may play a role in hemodynamic changes in retinal diseases in which hypoxia and changes in NO metabolism are involved in the disease pathogenesis. Vasomotion refers to spontaneous diameter oscillations in resistance vessels that facilitate blood flow, fluid homeostasis, and tissue oxygenation.1 A recent study has shown that the power of vasomotion in retinal arterioles is raised by an increase in the arterial blood pressure and is reduced by an increase in metabolic activity induced by exposure of the retina to flickering light.2 Increased metabolic activity results in a higher oxygen demand, and the resulting hypoxia induces dilatation of the resistance vessels that can increase perfusion and restore normoxia.3,4 Hypoxia-induced vasodilatation in vitro is mediated by a number of compounds, including nitric oxide (NO),5,6 and a previous study has also shown that inhibition of the synthesis of NO increases the frequency of vasomotion in porcine retinal arterioles in vitro.7 However, it is unknown whether these findings can be reproduced in human retinal arterioles in vivo. The goal of the present study was to examine spontaneous diameter changes of larger retinal arterioles and their impact on hydraulic conductance in two groups of healthy subjects before and during breathing of a hypoxic gas mixture. The procedure was repeated after peroral administration of the NO donor nitroglycerin in one group of test persons or intravenous infusion of the NO antagonist NG-monomethyl-L-arginine (L-NMMA) in another group of test persons. The harmonic content of diameter oscillations were extracted by Fourier analysis, and the effects of the interventions were studied on the amplitudes of diameter oscillations within the following four frequency intervals 1: 0.5/min–3.5/min, 2: 4/min–7.5/min, 3: 8/min–17.5/min, and 4: 18/min–30/min. Materials and Methods Study Persons Diameter recordings of retinal arterioles were analyzed from 40 healthy subjects (27 men and 13 women), aged (mean, range) 26.2 (21.3–55.9) years, with 20 subjects from each of two previously conducted studies of, respectively, the vasoactive effects of the NO-donor nitroglycerin8 and the NO-synthesis inhibitor L-NMMA.9 The examinations were performed on younger adults, which ensured that the elasticity of the arterial walls was sufficient to allow changes in the diameter to occur.10,11 The subjects had no known previous systemic or ocular disease, and an ophthalmologic examination, including measurement of visual acuity, slit-lamp examination, and fundus photography, was normal. The study was approved by the regional committee for scientific ethics and followed the tenets of the Helsinki Declaration, and all participants had given their oral and written consent to participate. Examinations The participants had undergone a standard ophthalmic examination, including measurement of best-corrected visual acuity (BCVA) using Early Treatment Diabetic Retinopathy Study (ETDRS) charts, IOP (Tonoref II; Nidek, Gamagori Aichi, Japan), and slit-lamp examination. The measured BCVA score was (mean ± SD) 94.25 ± 2.94 ETDRS letters, and the IOP was (mean ± SD) 13.91 ± 2.74 mm Hg. The pupils were dilated using phenylephrine 10% (SAD; Amgros I/S, Copenhagen, Denmark) and tropicamide 1% (Alcon, Geneva, Switzerland) eye drops, and systemic blood pressure was measured after a minimum of 5 minutes of rest using a cuff (705IT; Omron Healthcare, Kyoto, Japan) placed on the upper left arm. Recording of Vessel Diameters All examinations were performed on the left eye. The participant was seated in front of the Dynamic Vessel Analyzer (DVA), which consists of a fundus camera with a video recorder connected to a computer that allows continuous recordings of the retina.12 The participant was instructed to fixate a bar inside the viewing system so that the retina was displayed in real time on the computer monitor, where the examiner could mark the vessel segment to be measured. The most linear segment with a length of at least 200 µm (approximate mean = 315 µm; range, 200–601 µm) located on either the upper or the lower temporal arteriole within one-half and three disk diameters from the optic disk (approximate starting point: mean = 2377 µm; range, 1052–4263 µm from the disk margin) was chosen and marked for the recording of vessel diameters (Fig. 1). During an examination, the DVA software automatically aligned successive images to adjust for saccadic eye movements and interrupted the recording during larger eye movements or blinking. The software extracted the vessel diameter based on the contrast between the vessel and the surrounding retina, which could be done at a video frequency (25 Hz) for every 10 µm along the selected vessel segment. An image from a DVA examination with the vascular segment chosen for diameter measurements marked with a white line. The oblique white line to the right is the fixation bar, of which the end is superposed on the fovea. Experimental Protocol Figure 2 shows the experimental protocol, which had two phases. The study protocol with DVA examination blocks before and during hypoxia performed before (phase I) and during (phase II) interventions on NO metabolism. Phase I: Before intervention on NO metabolism: 1.DVA examination block 1.DVA examination block DVA examination block Each examination block lasted 840 seconds,13 of which only the first period of 120 seconds during rest was used in the present study. The systemic arterial blood pressure was measured at the end of the period using an oscillometer (705IT; Omron Healthcare) with the cuff on the upper left arm. 2.Hypoxia 2.Hypoxia Hypoxia Hypoxia was initiated by breathing a hypoxic gas mixture consisting of 12.5% oxygen and 87.5% nitrogen through either a face mask (Intersurgical, Berkshire, UK)9 or a silicon mouthpiece connected to a two-way nonrebreathing valve (Hans Rudolph Inc, Shawnee, KS, USA).8 The systemic oxygen saturation was monitored using a pulse oximeter (Masimo Rad-5v; Infiniti Medical, Greve, Denmark) placed on the middle finger of the left hand. After 10 minutes, a second DVA examination was performed while breathing of the hypoxic gas mixture was continued. Phase II: The two steps in phase I were repeated with the following modifications:Nitric oxide inhibition: Twenty participants received intravenous administration of the nonselective nitric oxide synthase (NOS) inhibitor L-NMMA (Bachem AG, Bubendorf, Switzerland) in a cubital vein of the right arm. The injection started with 6 mg/kg of the compound dissolved in saline injected over 5 minutes, followed by a constant infusion of the compound of 60 µg/kg/min. Both DVA examinations were performed during the continuous infusion of the compound.9Nitric oxide donor: In the 20 other participants, nitroglycerin 0.5 mg (DAK, Orifarm Healthcare, Odense, Denmark) was administered sublingually 5 minutes before repetition of the DVA examinations.8 Nitric oxide inhibition: Twenty participants received intravenous administration of the nonselective nitric oxide synthase (NOS) inhibitor L-NMMA (Bachem AG, Bubendorf, Switzerland) in a cubital vein of the right arm. The injection started with 6 mg/kg of the compound dissolved in saline injected over 5 minutes, followed by a constant infusion of the compound of 60 µg/kg/min. Both DVA examinations were performed during the continuous infusion of the compound.9 Nitric oxide inhibition: Twenty participants received intravenous administration of the nonselective nitric oxide synthase (NOS) inhibitor L-NMMA (Bachem AG, Bubendorf, Switzerland) in a cubital vein of the right arm. The injection started with 6 mg/kg of the compound dissolved in saline injected over 5 minutes, followed by a constant infusion of the compound of 60 µg/kg/min. Both DVA examinations were performed during the continuous infusion of the compound.9 Nitric oxide donor: In the 20 other participants, nitroglycerin 0.5 mg (DAK, Orifarm Healthcare, Odense, Denmark) was administered sublingually 5 minutes before repetition of the DVA examinations.8 Nitric oxide donor: In the 20 other participants, nitroglycerin 0.5 mg (DAK, Orifarm Healthcare, Odense, Denmark) was administered sublingually 5 minutes before repetition of the DVA examinations.8 Data Analysis The analyses of diameter changes were performed using a program in C++ (Code:Blocks open source vs 20.03) written by the second author. At each position on the vessel, the diameter values that were missing because of blinking (mean = 596; range, 59–1326) were substituted with the average of the remaining diameter values over time. For each position on the vascular segments, Fourier analysis was performed on the recordings of 120 seconds from each individual period, each corresponding to 3000 successive diameter measurements. The amplitudes were calculated for oscillation frequencies between 0.5/min and 30/min (1–60 harmonic oscillations), and the amplitudes were squared to powers to eliminate negative values. Subsequently, the amplitudes at each frequency were averaged for all the studied points along the vascular segment to represent one frequency pattern for this segment. The vascular resistance decreases with increasing vessel diameter to the power of 4. This implies that a given percentage increase in the diameter will increase the hydraulic conductance more than a similar percentage decrease in the diameter will reduce this conductance. Therefore, the hydraulic conductance in a vessel with an oscillating diameter will be higher than the conductance in a vessel with a constant diameter at the average of the oscillations. To study the effect of the diameter oscillations on hydraulic conductance, each diameter measurement obtained over time at each position on the vessel was raised to the power of 4, and the powered values were summed and averaged. The fourth root of this average of summed values was calculated to represent the diameter that a vessel with a uniform diameter with the same hydraulic conductance would have had. The percentage increase in this value from the average diameter of the vessel was subsequently calculated to represent the effect of the diameter oscillations on the hydraulic conductance. Statistical Analyses All tests were performed in Stata version 14.2 (StataCorp, College Station, TX, USA). For each of the two interventions, L-NMMA or nitroglycerin, the powers for each frequency in the Fourier analyses were considered before and during hypoxia before and during the drug intervention, resulting in eight groups of data. To identify frequency bands of interest, the powers of successive series of four harmonic oscillations (i.e., 1–4, 2–5, 3–6 . . . 57–60) were summed for each of the eight groups, and for each of the two drug interventions, the difference between these power sums before and during hypoxia was tested exploratively by paired t-tests. This revealed specific response patterns with significant effects of interventions among the groups for overlapping series of oscillations within the following four frequency intervals (harmonic oscillations): 1: 0.5/min–3.5/min (1–7), 2: 4/min–7.5/min (8–15), 3: 8/min–17.5/min (16–35), and 4: 18/min–30/min (36–60). For each person, the sum of the powers within these four bands was calculated for each experimental condition, and the differences between baseline, hypoxia, drug, and hypoxia combined with drug were tested with repeated-measures ANOVA. When the ANOVA showed a significant difference among the groups, the analysis was followed by t-tests among pairs of the groups to identify the background for the variation among the groups. Results The systemic arterial oxygen saturation at baseline in phase I was (mean ± SD) 98.8% ± 4.3%, which the breathing of a hypoxic gas mixture reduced significantly by (mean ± SD) 6.0% ± 3.8% (P < 0.0001). The response was not significantly different from that in phase II.8,9 The mean arterial pressure (MAP) was (mean ± SD) 86.4 ± 11.3 mm Hg at baseline. In phase II, the infusion of L-NMMA significantly increased the MAP by (mean ± SD) 13.9 ± 6.5 mm Hg (P < 0.0001), whereas administration of nitroglycerin significantly reduced the MAP by (mean ± SD) 10.7 ± 15.6 mm Hg (P < 0.0001). Figure 3 shows an example of a diameter recording before and during systemic hypoxia. The average arteriolar diameter was significantly increased (P = 0.04) from 120.3 ± 19.48 µm before hypoxia to 122.9 ± 19.26 µm during hypoxia. The average diameter was significantly (P = 0.005, paired t-test) reduced to 114.9 ± 22.2 µm by L-NMMA infusion but nonsignificantly (P = 0.84, paired t-test) increased to 116.3 ± 13.48 µm after the administration of nitroglycerin. Representative diameter traces during normoxia and following hypoxia in experimental phase I. It appears that hypoxia increases the diameter and the amplitude of spontaneous diameter oscillations. Figure 4 shows the mean and the 95% confidence interval of the powers of the studied harmonic oscillations at baseline from all 40 persons. For both drug interventions, there were significant differences (repeated-measures ANOVA, P < 0.0001 for all comparisons) between the four experimental conditions in all frequency bands. The power (mean, 95% confidence interval) of the harmonic oscillations at baseline. The vertical dashed lines delimit the four studied frequency bands. Figure 5 shows that in frequency band 1 (0.5/min–3.5/min), the change from normoxia to hypoxia increased the power of diameter oscillations significantly (P < 0.03 for both comparisons). The effect was preserved in the presence of L-NMMA (P < 0.02) but was eliminated by nitroglycerin (P = 0.82). In frequency band 2 (4/min–7.5/min), hypoxia had no significant effect on the power of the diameter oscillations, but L-NMMA significantly reduced the power (P < 0.01) and nitroglycerin significantly increased the power of the oscillations (P = 0.03). In frequency band 3 (8/min–17.5/min), hypoxia significantly increased the power of oscillations (P < 0.003 for all comparisons), but there was no significant effect of nitroglycerin or L-NMMA on the powers. In frequency band 4 (18/min–30/min), the power of the oscillations was significantly increased by L-NMMA (P < 0.001) and hypoxia (P < 0.006 for all comparisons), whereas nitroglycerin had no significant effect on the powers (P > 0.4 for all comparisons). The power (mean ± SEM) of the harmonic oscillations in each of the frequency bands 1 (0.5/min–3.5/min), 2 (4/min–7.5/min), 3 (8/min–17.5/min), and 4 (18/min–30/min). Asterisks indicate significant differences from baseline. The hydraulic conductance caused by the oscillating diameters at baseline was 1.34% ± 0.23%, and there were no significant differences in this parameter among any of the interventions (repeated-measures ANOVA, P > 0.09 for all comparisons). Discussion The present study has shown that hypoxia and interventions on NO metabolism can affect spontaneous diameter oscillations in retinal arterioles. The data were collected using the DVA, which has been shown to be suitable for the measurement of diameter changes of retinal arterioles in vivo with a high reproducibility.12 The recording of vessel diameters with a frequency of 25 times per second using this technique allowed the study of the harmonic content of diameter changes in retinal arterioles in vivo within a range of frequencies that can be assumed to have importance for mass transport, fluid homeostasis, and oxygenation.14 The number of studied persons in each group exceeded the 17 persons that had been found to be needed for the study of diameter changes in retinal vessels using the DVA.10 However, it is unknown whether this calculation of statistical power is also valid for the study of vasomotion, and therefore, it cannot be excluded that aspects of diameter regulation related to this phenomenon may have been overlooked. Additionally, an ideal experimental setup would have implied local interventions on oxygen saturation and NO metabolism, which was not feasible in vivo. Therefore, it cannot be excluded that the observed responses had been secondary to the effects of the systemic interventions on the extraocular vascular system.15 The local effects in the retina can be expected not to involve the autonomic nervous system, which is not represented in the retinal vasculature. This implies that the local responses had been strongly influenced by other mechanisms and signal molecules,16,17 of which oxygen is particularly important due to the high retinal metabolism.18,19 In cardiovascular research, amplitudes of harmonic oscillations are normally squared to result in powers that describe the effects of both increasing and decreasing diameters around the mean. The powers are summed for oscillations within certain frequency bands denoted very low frequency, low frequency, high frequency, very high frequency, or ultra-high frequency located within an overall frequency range from below 0.04 Hz to 12.5 Hz that are relevant for the heart and the larger systemic vessels.20–22 However, this categorization may not be relevant for smaller vessels and oscillatory phenomena occurring within or across the boundaries of these frequency bands. Therefore, in a stepwise procedure, a successive series of four harmonic oscillations were analyzed, representing frequencies from half to 30 periods per minute considered to be of potential physiologic relevance. This revealed four broader frequency bands (denoted 1–4) within which consistent changes in power could be observed in response to hypoxia or to one of the drug interventions. The fact that hypoxia increased the power of vasomotion in three of the four defined frequency bands confirms previous findings that hypoxia can enhance vasomotion23 and supports the hypothesis that vasomotion contributes to improving tissue oxygenation.24–26 However, it remains to be explained why hypoxia had no effect on the power in frequency band 3 (4/min and 7.5/min). This may be due to a lack of synchronization of the vascular supply with the metabolic need at these frequencies. The increasing amplitude of diameter oscillations at low and high frequencies will increase the hydraulic conductance of the vessels, and the resulting increase in blood flow may compensate for a reduced oxygen saturation in retinal vessels during hypoxia. However, the hypoxia induced in the present study corresponded to breathing at an altitude of approximately 3800 m within the normal physiologic range for a person acclimatized to this altitude. This may explain that the diameter oscillations of the retinal arterioles were insufficient to affect the hydraulic conductance of these vessels. It is possible that a further decrease in the oxygen saturation to a pathologic level may further increase the amplitude of vasomotion and the hydraulic conductance, with a resulting increase in the delivery of oxygen to the retinal tissue. The effects of NO supplementation and NO synthesis inhibition on retinal vasomotion reported in this study are new and support data from animal studies suggesting that NO can be involved in the regulation of vasomotion.1 These studies have shown that especially endothelium-derived NO may be important for synchronization of smooth muscle cells.27 In ex vivo studies on porcine retinal arterioles, it has also been shown that inhibition of NO synthesis can reduce the frequency of spontaneous diameter oscillations in porcine retinal arterioles at frequencies within the range of 0.64 to 0.97 oscillations per minute.7 These frequencies were located within frequency band 1 in this study, but the number of frequencies to be summed was too low and the variation of the powers too high to allow a comparison among the groups. In the previous in vitro experiments, inhibition of NO synthesis had no effect on the amplitude of the oscillations, which is in accordance with the present in vivo finding of a lack of effect of L-NMMA on the power of oscillations in frequency band 1. Previous in vitro studies have also shown that vasomotion and other modalities for retinal autoregulation can differ among smaller and larger retinal arterioles.4,28,29 This evidence could not be confirmed in the present study, which was limited to the analysis of arterioles with a diameter larger than 80 µm.12 The finding that the power of diameter oscillations with frequencies between 4/min and 7.5/min (frequency band 2) was increased by nitroglycerin and reduced by L-NMMA suggests that NO has a stimulating effect on spontaneous diameter oscillations within this frequency band. This was opposite to the observation at higher frequencies (band 4), where L-NMMA appeared to increase the power of the diameter oscillations. These findings support an effect of NO on the regulation of retinal vasomotion within specific frequency intervals and may suggest future targets of intervention on retinal vasomotion in health and disease. In conclusion, the study has shown that hypoxia and interventions on NO metabolism can change the amplitude of spontaneous diameter oscillations in retinal arterioles. This supports that vasomotion is involved in the regulation of retinal hemodynamics under normal conditions. Disturbances in this regulation may be involved in the development of retinal disease where hypoxia and changes in NO metabolism are involved in the disease pathogenesis.