Inhibition of cystathionine‐gamma lyase dampens vasoconstriction in mouse and human intracerebral arterioles

In extracerebral vascular beds cystathionine‐gamma lyase (CSE) activity plays a vasodilatory role but the role of this hydrogen sulfide (H2S) producing enzyme in the intracerebral arterioles remain poorly understood. We hypothesized a similar function in the intracerebral arterioles.


| INTRODUCTION
Hydrogen sulfide (H 2 S) is the third and most recently recognized member of the gasotransmitters family together with nitric oxide (NO) and carbon monoxide (CO). 1 H 2 S is endogenously synthesized by three different enzymes called cystathionine-gamma lyase (CSE), cystathioninebeta synthase (CBS), and 3-mercaptopyruvate sulfur transferase. 2 H 2 S production has been confirmed in the mouse brain cortex and in post-mortem material from the human brain stem 3,4 but the physiological role of the H 2 S producing enzymes in the mouse and human brain vasculature still remains to be fully understood. All three H 2 S-producing enzymes are expressed in the cardiovascular system, but CSE is the one mainly responsible for the majority of the endogenous H 2 S production in the vasculature. 5,6 Within the extracerebral artery wall, CSE is abundantly expressed both in the endothelium and in the vascular smooth muscle cells (VSMC), where it has been suggested to play a crucial role in the regulation of the vascular tone. [7][8][9][10] H 2 S has pleiotropic effects, acting both as a vasodilator and a vasoconstrictor, in various vascular beds such as the aorta and the mesenteric arteries. 7,10,11 Most of the studies emphasize its vasodilatory action like nitric oxide (NO) as seen in aorta, carotid-, coronary-, and mesenteric arteries. [12][13][14][15] However, it should be emphasized that such an interpretation of the biological functions of H 2 S is too simplistic. 16 The exact biological function of H 2 S depends on many factors such as ts relevant concentration, the type of vascular bed that it acts on, and the examined animal species. 17 In general, high concentrations of exogenously administrated H 2 S donors tend to induce vasodilation whereas low concentrations of endogenously or exogenously derived H 2 S tend to induce vasoconstriction. 18 In the brain vasculature CSE has been localized in the pial arterioles from pigs, where CSE was very abundant in the VSMC. 19 The biological effects of H 2 S in the cerebral vasculature have been investigated in the larger arteries in rodents where the effect of exogenously administrated H 2 S donors had a vasodilatory effect. [20][21][22][23] To the best of our knowledge, our study is the first investigating of the role of endogenous CSE-derived H 2 S in both human and mouse intracerebral arterioles. This type of vascular bed acts like a bridge between the larger diameter cerebral arteries and the brain capillaries. 24 Moreover, intracerebral arterioles play a crucial role in reassuring adequate blood supply of vulnerable brain areas and malfunction of this vascular bed, through for example extensive vasoconstriction, can increase the susceptibility of the brain to stroke-mediated ischemic injury and Alzheimer's disease. 25 Considering the abundant expression of CSE and the crucial role of CSE-derived H 2 S in regulating the vascular tone in other vascular beds, 18 the aim of our study was to test the hypothesis that CSE-derived H 2 S would contribute to vasodilatory mechanisms and that it would have similar biological functions as NO. First we investigated (i) whether CSE is expressed in human and mouse intracerebral arterioles and (ii) then we tested the role of CSE in the regulation of vascular tone in these arterioles. In contrast to the hypothesis, our results indicate that CSE-derived H 2 S supports the high K + -induced vasoconstriction but not the high K + -induced secondary dilatation. The K +induced secondary dilatation is supported by endothelial NO synthase (eNOS) and soluble guanylate cyclase (sGC) independently of CSE. 2 | RESULTS 2.1 | H 2 S producing enzymes are expressed in the human and mouse brain cortex and intracerebral arterioles CSE mRNA expression was detected in both human ( Figure 1A,B) and mouse ( Figure 1C,D) brain cortex and intracerebral arterioles by RT-PCR. For the human mRNA expression of CSE and the housekeeping gene RPL41 were consistently detected in brain cortex from both males and females, except in one male, while the CSE mRNA PAG on the K + -induced vasoconstriction in the mouse arterioles and attenuated the K + -induced secondary dilatation significantly.
Conclusion: CSE contributes to the K + -induced vasoconstriction via a mechanism involving H 2 S, eNOS, and sGC whereas the secondary dilatation is regulated by eNOS and sGC but not by CSE.

K E Y W O R D S
contractility, cystathionine-gamma lyase, endothelial NO synthase, hydrogen sulfide, intracerebral arterioles, soluble guanylate cyclase expression in matching intracerebral arterioles were more variable ( Figure 1A, B). For the mouse brain cortex and matching intracerebral cerebral arterioles, mRNA levels of CSE and housekeeping gene βactin were detected in all samples from both males and females ( Figure 1C, D). Also, another H 2 S producing enzyme, CBS, was expressed in all analyzed human ( Figure 1B) and mouse ( Figure 1D) intracerebral arteriole samples, again with some variation.

| Pharmacological inhibition of CSE with PAG in human intracerebral arterioles attenuates the K + -induced contractions
Administration of a high potassium solution (KPSS, 70 mM) in human intracerebral arterioles ex vivo induces a rapid contraction followed by a secondary dilatation, as shown in the representative micrographs of a human intracerebral arteriole in Figure 2A, B. Representative images of the maximum contraction in the presence of different pharmacological reagents (PAG 120 μΜ, PAG 120 μΜ + NaSH 10 μΜ) is presented in Figure 2A.
As shown in Figure 2B, 20 min pre-incubation with 'PAG' (120 μM, purple curve) attenuates the high K +induced maximum contractions. This attenuation of the constriction by PAG could be reversed in the presence of the H 2 S donor, NaSH (10 μΜ Figure 2C, dashed purple line showing PAG trace from Figure 2B for better visualization purposes). The mean maximal contractions are shown in Figure 2D, showing a significant reduction in the maximal K + -induced contractions by PAG, which was significantly reversed by NaSH. The actual baseline diameters before potassium stimulation in the three conditions did not differ significantly ( Figure S3A). In a separate subset of arterioles, we also examined the effects of NaSH (10 μΜ) without pre-incubation with PAG. As shown in the Figure S3B,C, exogenous NaSH did not have any significant effect on the K + -induced contractions (p value: 0.13).
Interestingly, we did not observe any significant effect of PAG on the K + -induced secondary dilatations in the human intracerebral arterioles over time or in terms of mean secondary dilatation for each experimental condition ( Figure 2B,E). Moreover, the H 2 S donor, NaSH (10 μΜ) did not have any effect on the secondary dilatations irrespective of whether it was administrated in the presence of PAG ( Figure 2C,E) or alone ( Figure S3D). Finally, the actual baseline diameters before potassium stimulation (μm) did not differ significantly for any of the tested experimental conditions (Control or NaSH alone) as shown in Figure S3E.

| Pharmacological inhibition of CSE with PAG in mouse intracerebral arterioles attenuate the K + -induced contractions
Representative micrographs of a mouse intracerebral arteriole under baseline conditions (baseline) and under maximum contraction in the presence of different pharmacological reagents (PAG 120 μΜ, PAG 120 μΜ + NaSH 10 μΜ) is presented in Figure 3A.
Like in isolated human arterioles, administration of a high potassium solution (KPSS, 70 mM) in mouse intracerebral arterioles ex vivo induces a rapid contraction followed by a secondary dilatation, as shown for the "Control" trace in Figure 3B. Pre-incubation with PAG (120 μΜ) attenuates the high K + -induced maximum contractions ( Figure 3B). We next pre-incubated the same vessel with PAG and added the H 2 S donor, NaSH (10 μΜ) which reversed the effect of PAG on the K + -induced maximum contraction Figure 3C (dashed purple line show the PAG trace from Figure 3B). The mean maximal contractions are shown in Figure 3D, showing a significant reduction in the maximal K + -induced constriction by PAG, that was significantly reversed by NaSH. The individual responses of male and female mice are depicted by red and black symbols, respectively, in all graphs. As shown in Figure S4A there were not any significant gender differences for all the "Control" and "PAG" treated arterioles. The actual baseline diameters before potassium stimulation in the three conditions did not differ significantly ( Figure S4B).
In a separate subset of mouse arterioles, we also examined the effects of NaSH (10 μΜ) without pre-incubation T A B L E 1 Human clinical information paraclinical data obtained at the day of surgery. with PAG. NaSH significantly attenuated the K + -induced contractions ( Figure S4C,D). Akin to the human results, we did not observe any significant effect of PAG on the K + -induced secondary dilatations in the mouse brain intracerebral arterioles either over time or in terms of mean secondary dilatation for each experimental condition ( Figure 3B,E). Moreover, the H 2 S donor, NaSH (10 μΜ) did not have any effect on the secondary dilatations irrespective of whether it was administrated in the presence of PAG ( Figure 3C,E) or alone ( Figure S4E). Finally, the actual baseline diameters before potassium stimulation (μm) did not differ significantly for any of the tested experimental conditions (Control or NaSH alone) as shown in Figure S4F.

| Pharmacological inhibition of eNOS with L-NAME attenuates the effect of PAG on K + -induced contractions and reduces the level of the secondary dilatations in mouse brain intracerebral arterioles
To clarify the underlying mechanism and to test if eNOS activity plays a role on the effect of PAG on the K + -induced contractions, we tested the effect of pre-incubation with the eNOS inhibitor L-NAME (100 μΜ, 20 min) in mouse intracerebral arterioles. L-NAME (blue line) administration over time led to a right-shift of the responses to high potassium solution over time compared to "Control" and significantly lowered the values of the % of mean arteriolar diameter from 75 to 85 s as indicated in Figure 4A. Administration of the same concentration of L-NAME in the presence of PAG (120 μM) led to significantly higher contraction compared to PAG alone ( Figure 4C). Administration of L-NAME irrespective of the presence or absence of PAG led to significantly reduced secondary dilatations compared to "Control" as indicated in Figure 4D. Baseline arteriolar diameter for all conditions was not significantly affected ( Figure S5A).

| Pharmacological inhibition of sGC with ODQ attenuates the effect of PAG on the K + -induced contractions and reduces the level of the secondary dilatations in mouse brain intracerebral arterioles
To further understand the underlying mechanism, and to test if sGC activity plays a role in the effect of PAG in the K + -induced contractions, we tested the effect of preincubation with the highly selective sGC inhibitor ODQ (10 μΜ, 20 min) in murine intracerebral arterioles. Even though ODQ (10 μM, blue line) alone did not have any significant effect on the K + -induced contractions over time ( Figure 5A) or mean maximal contraction ( Figure 5C), it significantly reversed the reduced K + -mediated contraction by PAG (gray line indicated as "ODQ + PAG" in Figure 5B,C).
As expected ODQ (10 μΜ) significantly reduced the K + -induced secondary dilatation both in the absence and after pre-incubation with PAG in murine arterioles as indicated in the graphs over time and for the secondary dilatation mean values of each group ( Figure 5A,B,D). Baseline arteriolar diameter for all conditions was not affected ( Figure S5B). A graphical representation of our findings and proposed mechanisms in relation to the effect of ODQ, L-NAME and PAG is given in Figure S6.

| DISCUSSION
In this study, we tested the hypothesis that H 2 S produced by CSE in the intracerebral arterioles contributed to the vasodilatory effects as seen in other extracerebral vascular beds. 5,11,29 Surprisingly, we found that, opposite to our original hypothesis, inhibition of CSE dampened the K + -induced vasoconstriction in both human and mouse intracerebral arterioles, while no effects were seen on the secondary dilatation after K + -mediated contraction. The effect of the CSE pharmacological inhibitor on vessel contractility could be reverted by an H 2 S donor NaSH and after pharmacological inhibition of NO and cGMP production.
In support of our findings, a recent study showed that aortas or carotid arteries from CSE KO mice have impaired contractile responses to phenylephrine. 30 Moreover, PAG can attenuate the prostaglandin F(2α) induced vasoconstrictions in rat pulmonary arteries. 31 PAG has been reported to have some off-target effects and inhibit also other PLP-dependent enzymes such as some aminotransferases, in higher mM-range concentrations. 32 Therefore, we cannot completely exclude off-target effects of PAG, at least in the human intracerebral arterioles that present a more variable expression of both CSE and CBS. However, we conclude that in our study most of the effects of PAG are mediated through CSE inhibition since (a) the effects of PAG are reversed by the H 2 S donor, NaSH, and (b) a relatively low concentration of PAG was used that is known to target only CSE and not CBS. 33,34 Interestingly, we found that when NaSH is administrated alone without the presence of PAG, it also dampens the K + -induced vasoconstrictions in mouse intracerebral arterioles whereas on the human arterioles a non-significant attenuating effect was observed. This species-specific difference could be of various reasons: (a) the human samples variate more in terms of age, gender, long term medications, genetic background, or other comorbidities. In support we did also find variable expression of arteriolar CSE mRNA and CBS mRNA in the human patients compared to mice, which could partly be F I G U R E 3 Pharmacological inhibition of cystathionine-gamma lyase attenuates the high potassium-induced contractions in mouse intracerebral arterioles. (A) Representative micrographs of a mouse intracerebral arteriole under baseline conditions, after high potassium induced maximal contraction, indicated as (K + ), after pre-incubation with L-propargylglycine (PAG) (120 μM, K + + PAG) and in the presence of the hydrogen sulfide (H 2 S) donor sodium hydrosulfide (NaSH) (10 μM) + PAG (120 μΜ), indicated as (K + + PAG + NaSH). The luminal diameter at the point of maximum contractions that was used for the analysis is indicated by a black line vertical line inside the vessel lumen and the exact size (in μm) of the inner diameter is given in brackets on the right side of each representative micrograph. Scalebar: 30 μm. (B) The % changes of the luminal diameter over time to high potassium before (Control) and after 20 min pre-incubation with PAG (120 μΜ, PAG) in perfused mouse intracerebral arterioles. (C) The % changes in the luminal diameter of mouse intracerebral arterioles to high potassium solution in the presence of the H 2 S donor NaSH (10 μM, PAG + NaSH). The dashed purple line shows the trace of PAG in (B). (D) The % changes of the maximum contraction to the high potassium solution (Control), 20 min after pre-incubation with PAG (120 μΜ, PAG) alone, or presence of the H 2 S donor NaSH (10 μM, PAG + NaSH). (E) The % changes of the secondary dilatation to the high potassium solution (Control), after 20 min of pre-incubation with PAG (120 μΜ, PAG) alone and in the presence of the H 2 S donor NaSH (10 μΜ, PAG + NaSH). Data are analyzed with repeated measures two-way ANOVA (B, C) and one-way repeated measures ANOVA (D, E), both test with Bonferroni post-test and presented as mean ± SEM, n = 6/group for (A-D), red symbols: Arterioles from male mice, black symbols: Arterioles from female mice. All three traced conditions "Control", "PAG", "PAG + NaSH" in (B, C) are tested on the same arteriole but for optimal visualization graphs in figure (B, C) have been separated. The dashed horizontal lines in (B, C) indicate the level of % changes in arteriolar diameter at 60% and 80%. *p < 0.05, **p < 0.01. because of poor RNA quality of the more extended procedure to obtain and isolate the human intracerebral arterioles compared to mice or, (b) the murine arterioles are more susceptible to changes in H 2 S compared to humans. In the limited studies investigating the effect of either exogenous or endogenous H 2 S in human vascular beds, it seems that higher concentrations of H 2 S donors, compared to the ones needed for mice, are required for achieving a significant vasodilatory effect. 35,36 To explore the underlying mechanism by which CSE supports the K + -induced vasoconstrictions in murine intracerebral arterioles, we investigated the eNOS/NO/ sGC/cGMP pathway since a recent study showed that CSE-derived H 2 S-induced contractions require an intact eNOS/NO/sGC/cGMP and occur in an endothelial celldependent pathway manner that involves cIMP. 30 This pathway was observed in aortic ring preparations from elder mice. 30 The interaction between NO and H 2 S is complex. Some studies have shown that H 2 S can increase NO production and/or activity mediated by post-translational activation of eNOS or via phosphodiesterase inhibition leading to vasodilation. [37][38][39][40] While other studies indicate that CSE inhibits the eNOS-derived NO and supports vasoconstriction in vascular beds such as the aorta, the carotid artery, and the coronary arteries. 30,41,42 Here we show that neither L-NAME (eNOS inhibitor) nor ODQ (sGC inhibitor) have any significant effect on the K + -induced contractions, when they are administrated alone. We have previous seen the lack of effect of L-NAME on potassium-mediated constriction, 28 but it is the first time we tested here the effect of ODQ. Our results indicate that eNOS and sGC activity cannot directly attenuate the K + -induced contractions that is somewhat of an unexpected finding. The intracerebral arterioles are small in diameter and with a quite distinct and different anatomy and physiology (i.e., the presence of the blood brain barrier and perivascular astrocytes) compared to extracerebral arteries. It is well recognized that regulation of vasoreactivity between intracerebral arterioles and extracerebral arterioles may differ 43,44 and thus the response to NO and/or H 2 S. Some other possible explanations for this result might have to do with (a) the vessel diameter, (b) the concentration of K + , and (c) with the fact that CSE might hinder eNOS activity under baseline conditions.
Larger diameter middle cerebral arteries treated with L-NAME have a reduced baseline diameter 45 which we, however, did not observe here with the smaller diameter intracerebral arterioles. Of note, small diameter renal afferent arterioles pretreated with L-NAME, have shown increased contractility when a wide range of K + concentrations was tested. 46 Here, a specific K + concentration was used based on the optimal conditions for vessels' response and viability but we cannot exclude the possibility that L-NAME or ODQ could have an effect in other K + concentrations.
Finally, CSE might play a critical role since both L-NAME and ODQ affect K + -induced contractions in the presence of the CSE inhibitor PAG. We show that CSE inhibition leads to attenuated contractions and this effect is reversed by inhibition of the eNOS/sGC axis by L-NAME or ODQ. It has been previously shown in mesenteric arteries, that CSE derived H 2 S, can scavenge the eNOSderived NO. 42 Thus, a proposed mechanistic explanation could be that the CSE-derived H 2 S, inhibits the eNOS/ sGC axis, and in that way supports the K + -induced vasoconstriction. Whether CSE-derived H 2 S can scavenge the eNOS-derived NO in the intracerebral arterioles should be further supported by future studies.
It is well established that eNOS induced NO activate sGC and thus increase intracellular cGMP and vasodilatation. 47 Previous studies from us have also shown that eNOS is necessary for mediating the high K + -induced secondary dilatation in mouse intracerebral arterioles. 28 Our present findings show that both L-NAME and ODQ attenuate the K + -induced secondary dilatations and therefore not only we have confirmed our previous findings on the role of eNOS in secondary dilatation, but we have also shown that sGC which is activated by NO, plays an important role in supporting the K + -induced secondary dilatations.
From the clinical point of view, it has been shown that there is a positive correlation between serum cysteine levels and poor clinical outcome in the ischemic stroke patients which could be correlated in vivo with enhanced activity of H 2 S synthesizing enzymes like CSE. 4,48 Our results from both human and mouse intracerebral arterioles indicate that the endogenous CSE contributes to the K + -induced vasoconstriction via an H 2 S-dependent mechanism. Therefore, our findings might have implications and potentially highlight the usefulness of CSE pharmacological inhibitors in cerebrovascular diseases with pathologically enhanced intracerebral arteriolar contractility such as in ischemic stroke. 49

| Human brain tissue
The use of human material was approved by the Regional Ethical Committee in the Region of Southern Denmark (S-20130048 and S-1300854) and was performed in agreement with the Helsinki Declaration. All the patients had given their written consent to participate in the study. Further details regarding the brain biopsies alongside with a detailed list of patient characteristics (i.e., gender, age, blood pressure, smoking history, and diagnosis) are listed in Table 1.

| Animals
This study was conducted in accordance with the Danish Animal Experiments Inspectorate (2020-15-0201-00499) under the Danish Ministry of Justice and followed the ARRIVE guidelines. 26 Murine studies were conducted in wild type C57BL/6J mice bought at 8 weeks old from Janvier laboratories (Le Genest-Saint-Isle, France) or local bred and housed at the Biomedical Laboratory at University of Southern Denmark with free access to standard chow and tap water until experiments (9-11 months old). In all experiments, an equal number of male and female mice were used to avoid sex bias. Further details are given in the Supplementary Section of Materials and Methods.

| Isolation of human or mouse intracerebral arterioles from human biopsies or mouse brains
Intracerebral arterioles from human and mouse were isolated with a microdissection procedure using forceps, under a stereomicroscope at 4°C and placed in appropriate preservation ATM medium (details of F I G U R E 4 Pharmacological inhibition of endothelial nitric oxide synthase (eNOS) reverses the effect of the cystathionine-gamma lyase (CSE) pharmacological inhibitor L-propargylglycine (PAG) on the high potassium-induced contractions, while it reduces the secondary dilatation by a CSE-independent mechanism. (A) The % changes of the luminal diameter over time to high potassium before (Control) and after 20 min pre-incubation with L-NAME (100 μΜ, L-NAME) in perfused mouse intracerebral arterioles. (B) The % changes of the luminal diameter to high potassium after 20 min incubation with PAG (120 μΜ) or 20 min pre-incubation of PAG (120 μΜ) in the presence of L-NAME (100 μM, PAG + L-NAME). The same arteriole was used in the experiments in both figure (A, B) but are graphical separated for optimal visualization. (C) The % changes of the maximum contraction to the high potassium solution (Control), after pre-incubation with L-NAME (100 μΜ, L-NAME), after pre-incubation with PAG (120 μΜ, PAG), and pre-incubation with PAG (120 μM) + L-NAME (100 μM, PAG + L-NAME). (D) The % changes of the secondary arteriolar dilatation to the high potassium solution (Control), after pre-incubation with L-NAME (100 μΜ, L-NAME), after pre-incubation with PAG (120 μΜ, PAG), and after pre-incubation with PAG (120 μM) + L-NAME (100 μM), indicated as (PAG + L-NAME). Data are analyzed with repeated measures two-way ANOVA (A, B) and one-way repeated measures ANOVA (C, D), both with Bonferroni post hoc test and presented as mean ± SEM, n = 6/group, red symbols: Male mice, black symbols: Female mice, The dashed horizontal lines in (C, D) indicate the level of % changes in arteriolar diameter at 60% and 80%. *p < 0.05, **p < 0.01, ***p < 0.001. medium composition and of the protocol are given in the Supplementary Section of Materials and Methods).

| Perfusion of human and mouse intracerebral arterioles
The perfusion of isolated intracerebral arterioles from humans and mice was performed as previously described in a custom-made system of mounting and perfusion pipettes in a thermostatic chamber. 27,28 The arteriolar diameter over time is expressed as % changes in diameter compared to the baseline (resting diameter) of each individual vessel. All experimental protocols started with an equilibration period (20 min) after the perfusion was established. Viability was tested by administration of a 70 mM KCl solution to the organ chamber. Details of the protocol used are given in the Supplementary Section of Materials and Methods. A graphical representation of the method is provided in the Figure S1.

| Reverse-transcription polymerase chain reaction (RT-PCR)
Total RNA was isolated using TriZol reagent (Invitrogen) from human and mouse cerebral cortex and isolated intracerebral arterioles. The RNA was then reversed transcribed using iScript cDNA synthesis kit (1708891, BioRad). PCR amplification protocol and primer sequences for human and mouse CSE alongside with the respective housekeeping genes are described in the Supplementary Section of Materials and Methods and in Table S1.

| Statistics
Data were analyzed with the help of Image J, Excel, and GraphPad Prism 9.0 software. The % changes in diameter between two groups over time were compared with paired t-test on each row followed by Holm-Sidak's multiple comparison test. The % changes in diameter between three groups or four groups over time were compared with repeated measures two-way ANOVA followed by Bonferroni multiple comparison test. The % of maximum contraction and % secondary dilatation data were analyzed by paired Student's t-test for comparisons between two groups or by one-way repeated measures ANOVA followed by Bonferroni multiple comparison test for comparisons of three or more groups. Sphericity was tested with Mauchly's test (https://www.stats kingd om.com/repea ted-anova -calcu lator.html) and it was assumed only in the cases where the p value from the Mauchly's test was lower or equal to 0.05. All the examined data passed the sphericity test, except from the data presented in Figure 4D (secondary dilatation in the presence of L-NAME). *Indicates significant differences with p < 0.05 as p < 0.01 as ** and p < 0.001 as ***. All data are presented in the graphs as mean ± SEM.