Gut microbiota and neuroinflammation in pathogenesis of hypertension: A potential role for hydrogen sulfide

Abstract

Inflammation and gut dysbiosis are hallmarks of hypertension (HTN). Hydrogen sulfide (H₂S) is an important freely diffusing molecule that modulates the function of neural, cardiovascular and immune systems, and circulating levels of H₂S are reduced in animals and humans with HTN. While most research to date has focused on H₂S produced endogenously by the host, H₂S is also produced by the gut bacteria and may affect the host homeostasis. Here, we review an association between neuroinflammation and gut dysbiosis in HTN, with special emphasis on a potential role of H₂S in this interplay.

Graphical Abstract

1. Evidence of neuroinflammation in hypertension

Hypertension (HTN) is a major risk factor for cardiovascular and cerebrovascular events and chronic kidney disease leading to high levels of mortality [1, 2]. The estimated prevalence of HTN among U.S. adults aged ≥20 is reportedly 46% [3]. Approximately 13.7% of hypertensive patients remain resistant or refractory to available anti-hypertensive treatments [4], which may in part be due to neurogenic mechanisms reflecting in increased sympathetic activity and reduced vagal cardiac tone [5, 6]. However, the mechanisms contributing to treatment-resistant HTN are still not completely understood [7]. Dysregulation of renin-angiotensin system (RAS) [8] and presence of inflammation in autonomic brain regions are touted as important contributors to the sympathetic overactivity in HTN [9, 10].

Neuroinflammation is an inflammatory response of central nervous system (CNS) mediated by cytokines, chemokines, reactive oxygen species (ROS), and secondary messengers produced by the resident CNS glia, endothelial cells, and peripherally-derived immune cells [11]. Neuroinflammation has been reported in HTN, however, it is still not clear whether it is a cause or a consequence of HTN, as RAS overactivity, blood brain barrier (BBB) disruption, peripheral immune system (IS) activation, and oxidative stress among other factors are all implicated in pathogenesis of HTN [8]. Correlation of HTN with serious neurological disorders such as Alzheimer’s disease and Parkinson’s disease, both characterized by a neuroinflammatory phenotype and degrees of autonomic dysfunction and blood pressure (BP) abnormalities, highlights the importance of adequate anti-hypertensive treatments for overall brain health [12–14].

Brain renin-angiotensin system, sympathetic outflow, neuroinflammation and hypertension

Numerous neurotransmitters and neuromodulators are involved in regulation of sympathetic outflow and neuroinflammation. A plethora of evidence implicates hyperactivity of brain RAS in development and maintenance of HTN in several experimental and genetic animal models including the spontaneously hypertensive rats (SHR) [15, 16], renin transgenic hypertensive rats [17, 18], Dahl salt-sensitive rats [19–21], deoxycorticosterone acetate (DOCA) salt-treated rats [22–24] and renal hypertensive rats [25–28]. Indeed, angiotensin (Ang) II levels were found to be high in cerebrospinal fluid of SHR [29] and in dogs with renal HTN [30], as well as in hypertensive humans following sodium depletion [31].

Within the CNS, the Ang II precursor angiotensinogen is reportedly synthesized in neuronal cells, glial cells and astrocytes [32–34]. Pro-hypertensive and pro-sympathetic action of brain Ang II is primarily mediated via the neuronal Ang type 1 receptor (AT1R) [35, 36]. The evidence for this is significant. For example, removal of AT1R in catecholaminergic neurons delayed the onset of Ang II-HTN and attenuated the sympathetic activation [37]. In addition, genetic inhibition of AT1R subtype in the paraventricular nucleus (PVN) of the hypothalamus decreased Ang II HTN [38], and decreased BP, sympathetic activity, and plasma Ang II and norepinephrine (NE) levels in the SHR [39]. AT1R is densely distributed in the ventrolateral medulla, nucleus tractus solitarii (NTS), lamina terminalis, supraoptic nuclei and the PVN, all known to involved in cardioregulation and regulation of body fluid and electrolyte balance [32]. Central AT1R-mediated actions of Ang II are reportedly mediated by phosphoinositide 3-kinase [40, 41], nicotinamide adenine dinucleotide phosphate (NADPH) oxidase [42–46], catecholamines [47–54], serotonin [54, 55], gamma-amino butyric acid (GABA) [56–60], glutamate [61, 62], acetylcholine [63], substance P [64–67], endothelial nitric oxide (NO) synthase and NO [68, 69], and vasopressin [70–72], all reportedly involved in modulating sympathetic activity. In line with this, PVN infusion of Ang II increased glutamate and decreased GABA levels in PVN, while elevating the BP [73]. Microinjection of Ang II into the PVN increased renal sympathetic nerve discharge, BP [74–76] and NO release [74], while microinjection of an AT1R antagonist Losartan reduced GABA_A receptor-mediated increase in renal sympathetic nerve activity and BP [77]. Moreover, the increase in BP in response to AT1R activation in the PVN of hypertensive Dahl rats is also suggested to be mediated by enhanced local glutamate receptor activation [78], possibly the glutamate N-methyl-D-aspartate (NMDA) receptor activation [79].

In addition to direct neuronal activation, Ang II is shown to mediate overexpression of brain pro-inflammatory cytokines (PIC) which may also contribute to sympathetic overactivity in HTN [73]. Several studies have reported a mild inflammatory state in several cardioregulatory brain regions in animal models of HTN (Table 1). However, direct effects of neuroinflammation on autonomic nervous system and BP in HTN remain unexplored. Early studies in the NTS of SHR suggested that elevated expression of junctional adhesion molecule-1 mediated the neuroimmune interaction in regulation of BP [80]. Another study showed that an injection of PIC interleukin (IL)-1β into the PVN of the hypothalamus increased BP, while overexpression of the anti-inflammatory IL-10 within the PVN reduced BP in Ang II rodent HTN [81]. Moreover, PVN microinjection of tumor necrosis factor (TNF)-α and IL-1β increased BP and sympathetic outflow in the SHR [82] suggesting direct neuronal effects, while elevated levels of IL-1β and decreased levels of IL-10 and nuclear factor kappa B (NF-κB) activity in the PVN have also been linked to HTN-induced sympathetic hyperactivity in Dahl salt-sensitive hypertensive rats [83, 84].

Table 1.

Presence of inflammation in the cardioregulatory brain regions in animals with hypertension.

HTN model	Species	Region	Reference
SHR	rat	PVN	[85–89]
		RVLM	[86, 90–94]
		NTS	[95–97]
2K1C	rat	PVN	[98]
		RVLM	[99]
Ang II-induced HTN	rat	PVN	[73, 85, 100–104]
		RVLM	[93]
High salt HTN	rat	PVN	[83, 84, 105–110]
		RVLM	[111, 112]
Stress-induced HTN	rat	PVN	[113]
		RVLM	[114]

Ang II: Angiotensin II; HTN: Hypertension; NTS: Nucleus tractus solitarii; PVN: Paraventricular nucleus; RVLM: Rostral ventrolateral medulla; SHR: Spontaneously hypertensive rat; 2K1C: Two-kidney one-clip

One potential source of increased PIC in the brain are the resident microglia, activation of which has been reported in HTN. Microglia are the resident immune cells of the brain involved in the maintenance of neural environment [115]. They are widely distributed but appear to have varied roles in specific brain regions [116, 117] playing an active role in both physiological and pathological states [118]. Microglia are activated upon various noxious stimuli and inflammatory states [119], upon which they undergo morphological changes [120] and release PIC [121] such as cytokines and prostaglandins, which can influence and modulate neuronal function during both physiological and pathological processes [122–124]. Expression of AT1R in has also been reported in glial cells [125–128]. It has been suggested that Ang II can increase the production and/or release of PIC from glia [7, 81, 100], which in turn increases the production of ROS that may increase neuronal discharge, thereby indirectly contributing to the sympathetic outflow and increased BP [129, 130]. Ang II binding to the AT1R can also stimulate the microglial NADPH oxidase, leading to oxidative stress and resulting in neuroinflammation [125, 126]. Activation of microglial RhoA/Rho kinase pathway is suggested to be involved in the Ang II-induced microglial activation [131–133]. However, AT1R expression has only been reported in activated microglia in certain neurological conditions [134]. Moreover, microglial activation in rodent models of HTN in which Ang II has been infused systemically [7, 81, 100] raises an important question whether Ang II acts directly or indirectly on microglia in HTN. Thus, the exact mechanisms of how Ang II induces microglial activation need to be clarified.

Alongside microglia, astrocytes are involved in neuroinflammation and are capable of producing PIC in response to various pro-inflammatory stimuli [135, 136]. Astrocytes express AT1R and can mediate Ang II-dependent increase in BP and sympathetic outflow in the PVN via inhibition of the astrocyte glutamate transporters [137]. Moreover, Ang II stimulates the production of IL-6 in SHR astrocytes in culture via the AT1R/ NF-κB/ROS pathway [138], while application of Ang II to human astrocytes in vitro increased the translocation of cytosolic components of NADPH oxidase to the cell membrane [139]. In neurons co-cultured with astrocytes, the Ang II-induced neuronal damage was inhibited by an AT1R blocker Valsartan, suggesting that astrocytes mediated central effects of Ang II [140]. In addition, microglia-astroglia communication is enhanced following microglial activation, which can indirectly affect neuronal function through strengthening of gliotransmission mediated by astrocytes [141]. For instance, TNF-α released by activated microglia reportedly potentiates glutamate release from the neighboring astrocytes, which can modulate synaptic plasticity and even lead to neurotoxicity [142]. In addition, microglial adenosine triphosphate is also shown to induce glutamate release by astrocytes which can excite proximal neurons [143].

Microglia can reportedly support adaptive synaptic plasticity through release of neurotrophic factors such as the brain-derived neurotrophic factors [144], while also being able to phagocytose synapses during neurodevelopment [145]. In line with this, minocycline, a microglial inhibitor with antibiotic and anti-inflammatory properties that can cross the BBB, can inhibit microglia in the PVN, thus reducing PIC and ROS production and alleviating HTN [81, 85, 146–149]. Concurrent central infusion of minocycline also attenuated increased expression of a microglial marker CD11b in lamina terminalis of rats receiving Ang II [150]. However, although microglial activation in the PVN was increased in Ang II HTN and in stroke-prone SHR, systemic or central administration of minocycline into the PVN failed to decrease BP [148, 151], possibly related to the animal model and dose/route of administration. Minocycline has also been shown to decrease BP in patients with treatment-resistant HTN, owing to its anti-inflammatory properties [152]. Moreover, central administration of a chemically modified tetracycline (CMT-3) with anti-inflammatory activity reduced microglial activation in the PVN in Ang II-induced rodent HTN [153]. Considering the above studies, agents with central anti-oxidant and anti-inflammatory properties may be beneficial in neurogenic HTN. In addition, centrally acting angiotensin converting enzyme (ACE) inhibitors [154–156] and AT1R blockers [156–160] may be useful in curbing neuroinflammation in HTN. However, the glial-neuronal communication in HTN is complex, and more sophisticated techniques are necessary to delineate the precise cellular and molecular mechanisms.

2. Correlation between gut dysbiosis and neuroinflammation in hypertension Gut microbiota

Gastrointestinal (GI) tract is a major interface between the host and external factors such as food, medication, alcohol and pathogens [161]. Mucosal surfaces of the GI tract are colonized by diverse microorganisms (bacteria, viruses and eukaryotes) collectively known as the ‘gut microbiota’, suggested to play significant roles in health and disease [161]. Gut microbiota is individually shaped by numerous factors such as the genetics, gender, age, IS, diet, drugs etc. [162]. The healthy adult gut microbiota consists of more than 100 trillion microorganisms [161], and Bacteroidetes and Firmicutes are the most abundant phyla [163].

Microbial colonization of the GI tract starts before birth and continues with the transfer of microorganisms during the delivery and breast-feeding, shaping the development of the host IS [164]. Gut microbiota has many beneficial functions in the body. It is involved in the conversion of food and indigestible components such as plant polysaccharides into nutrients such as short chain fatty acids (SCFA), which predominantly consist of acetic acid, butyric acid, and propionic acid, for the more efficient absorption and utilization by the host [165]; development of the intestinal epithelium and protection against tissue injury [166, 167]; regulation of fat storage [168]; modulation of enteric nervous system (ENS) [169]; development and maintenance of host IS [164]; and shape CNS development and function [170–172].

Possible association between gut dysbiosis and neuroinflammatory responses in hypertension

Although components of the microbiota are often referred as commensals, interaction between microbiota and the host includes mutualistic, parasitic and commensal relationships [173]. Mounting evidence indicates that any alteration in the balance between gut microbiota and mucosal IS resulting from diet, medication, pathogens etc., results in changes in balance and composition of intestinal commensal communities i.e. gut dysbiosis, which has been associated with a plethora of diseases and conditions such as the inflammatory bowel disease, diabetes, obesity, arthritis, neurodegenerative diseases and HTN, among others. [146, 174, 175].

In recent years the concept of the “gut-brain axis” has been introduced. Communication between the gut and the brain is bidirectional and includes neural pathways via the ENS, autonomic and spinal nerves, and humoral pathways via cytokines, hormones, and various neuropeptide signaling molecules as well as metabolic factors such as the SCFA and tryptophan [176, 177]. Evidence suggests a role for gut dysbiosis in several neuroinflammatory diseases [178–181]. Recently, gut dysbiosis has been reported in HTN [146, 182–189] and played an important role in BP control. Administration of SCFA causes vasodilation in both rodents and humans [190–193], and gut bacteria and/or their metabolites may affect the regulation of the circulatory system via ENS and the afferent sensory feedback to the brain that may modulate the activity of the cardioregulatory brain regions [194, 195]. In rodent models of HTN characterized by neuroinflammation, elevated BP is linked to increased gut permeability, heightened inflammatory state and increased sympathetic neuronal communication between the PVN and the gut [153, 188]. On the other hand, inhibition of microglial activation in the PVN decreased sympathetic activity, impacted selective gut microbial communities, attenuated gut wall pathology, and lowered BP in the Ang II-dependent rodent HTN [153]. Additionally, gut dysbosis may directly affect the IS and vice versa [196–199], while SCFA can dampen glial inflammatory responses [200–203] and in this way may be regulating BP [195, 204]. However, whether gut dysbiosis is a cause or a consequence of HTN, as well as the precise mechanisms that lead to development of gut dysbiosis in HTN and the impact of established gut dysbiosis on BP and neuroinflammation in HTN, is not known. Recent studies offered a clue to the possible mechanisms of gut dysbiosis-linked HTN, as the gut microbiota transplant from SHR to normotensive rats increased BP and sympathetic drive via activation of both systemic and central IS responses [197, 205, 206], suggesting that the microbiota-microglial crosstalk is important in regulation of BP.

Gut microbiota and microglia crosstalk

The gut microbiota can significantly affect microglia from before birth and throughout the adulthood, as the microbial-derived metabolites and neurotransmitters can regulate the inflammatory responses mediated by microglia in the CNS [200–203, 207–209]. The gut microbiota has been shown to affect microglia through microbial-derived metabolites and neurotransmitters [200–203, 207–209]. A study investigating interactions between the gut microbiota and microglia showed that germ-free mice displayed microglial immaturity and defects in microglial proportions, which are normalized by treatment with microbial SCFA [210]. In addition, an anti-inflammatory SCFA butyrate has been shown to reduce lipopolysaccharide (LPS)-induced responses in rat primary microglia [200–203, 211] and dampen inflammatory mediators in SHR primary astrocytes [200]. Since animal [146, 212] and human studies [213] have shown a decrease in butyrate-producing bacterial population in HTN with reduced plasma butyrate levels [195, 213], which may contribute to microglial activation we see in HTN.

Gut dysbiosis may also contribute to neuroinflammatory responses by increasing both gut and BBB permeability thus allowing infiltration of peripheral immune cells and PIC from systemic circulation into the brain [214–216]. Indeed, in rodent models of HTN, activation of peripheral IS is linked to microglial activation [85, 92], while elevated BP is marked by increased gut permeability, heightened inflammation, and increased sympathetic neuronal communication between the PVN and the gut [153, 188].

The vagus nerve is a crucial physiological component of gut-brain axis as well as the immune function [217, 218]. Efferent vagal branches innervate visceral tissues such as the liver, GI tract, thymus and spleen involved in regulation of immune and metabolic responses, while the afferent vagal branches report on the inflammatory and metabolic status in the periphery via their sensory afferent feedback to the NTS [219, 220]. In the gut, the vagus can sense the inflammatory mediators and relay these signals to the brain, and this may in turn activate vagal efferent responses to attenuate inflammation [221] in a process termed the vagal anti-inflammatory reflex. For example, intraperitoneal injection of microbial-derived LPS will cause a release of IL-1β which activates the vagal afferents in the periphery and, via the NTS, modulates efferent autonomic responses in order to curtail the inflammation in the periphery [222]. LPS has also been shown to cause glia-mediated neuroinflammation, while electric vagal nerve stimulation can dampen the LPS-induced activation of microglia and PIC production in brain [223]. In addition, gut bacterial-derived metabolites and neuroactive factors can directly activate the vagal afferents via receptors located on vagal nerve endings and on the cell soma [219, 224]. These studies support the role of the vagus nerve in gut-brain axis and regulation of neuroinflammatory responses.

3. Hydrogen sulfide in hypertension and neuroinflammation

Hydrogen sulfide (H₂S), traditionally known as a toxic gas with the smell of rotten eggs, is an endogenously produced signaling molecule in bacteria, plants, and animals including the mammals [225, 226]. Studies in animals and humans have shown the H₂S to be involved in diverse physiological and pathophysiological processes, including immune conditions, HTN, neurodegenerative diseases and metabolic disorders [227].

Hydrogen sulfide in human and rodent hypertension

H₂S is considered an endogenous vasoactive factor [222] that can elicit concentration-dependent vasorelaxation [223–226]. Several mechanisms have been proposed to contribute to this effect. H₂S can induce vascular smooth muscle relaxation through the activation of K_ATP channels leading to membrane hyperpolarization [228]. Moreover, H₂S has been suggested to dilate blood vessels [229] by promoting the release of NO [230], while NO can also enhance H₂S production from vascular tissues [228]. H₂S can also produce vasodilation by increasing intracellular cyclic guanosine monophosphate levels via inhibition of the phosphodiesterase activity [231]. In addition, H₂S can reportedly attenuate vascular inflammation [232], reduce ROS production [233], and inhibit the synthesis and release of renin [234] as well as lower ACE activity [235]. In this way, H₂S can produce BP-lowering effects by modulation of neural and immune activity. However, the full extent of H₂S BP-lowering effects remains unknown.

Mounting evidence suggests that H₂S may play a role in development of rodent and human HTN (Table 2). In the SHR, low plasma H₂S and diminished activity and expression of H₂S-producing enzyme cystathionine γ-lyase (CSE) has been reported, whereas intraperitoneal administration of sodium hydrosulfide (NaHS), an H₂S donor, alleviated HTN in the SHR by upregulation of the diminished H₂S synthase activity [236, 237] by increasing renal H₂S production and NO bioavailability and inhibiting renal RAS [238]. In the 2K1C (2kidney-1clip) hypertensive rats, plasma H₂S levels, CSE expression levels and BP were all normalized by the NaHS treatment [239, 240]. In Dahl salt-sensitive rats, high salt intake reduced H₂S in serum and the kidney and inhibited renal expression of H₂S-producing enzyme cystathionine β-synthase (CBS), while administration of NaHS normalized H₂S and inhibited renal RAS, thus preventing salt-sensitive HTN in Dahl rats [241]. In rats with hypoxic pulmonary HTN, low plasma H₂S levels were linked with decreased CSE activity in the lung, which was reversed by administration of an H₂S donor [242]. In the CSE knock-out mice, the developing HTN is linked with low serum H₂S levels, and injection of NaHS lowered BP in the 10-week-old CSE knock-out mice [243]. Chronic Ang II infusion increased BP, lowered plasma H₂S levels and downregulated kidney expression of CBS and CSE in mice, which were all reversed upon treatment with an H₂S donor GYY4137 [244]. Chronic administration of CSE and CBS enzyme inhibitors to normotensive rats resulted in a decrease in urinary excretion rate of sulfate, which is considered an indicator for endogenous H₂S production. Decrease in the excretion rate of sulfate was also associated with increased mean arterial pressure following dual enzyme inhibition [245].

Table 2.

Plasma/serum hydrogen sulfide levels in animals and patients with hypertension.

	Animal studies	Species	H2S levels	Reference
Animal model	SHR	rat	Low	[236, 237]
	Diabetic SHR	rat	Low	[252]
	2K1C	rat	Low	[239, 240]
	Dahl	rat	Low	[241]
	Hypoxic pulmonary HTN	rat	Low	[242]
	Load induced HTN	rat	Low	[251]
	CSE knock-out	mice	Low	[243]
	Ang II induced HTN	mice	Low	[244]
	Human studies		H2S levels	Reference
Type of HTN	Essential		Low	[240, 246, 247]
	Portal		Low	[248]
	Pulmonary		Low	[249]
	Preeclampsia		Low	[250]
	Lead-induced HTN		Low	[251]

Ang II: Angiotensin II; HTN: Hypertension; H₂S: Hydrogen sulfide; CSE: cystathionine γ-lyase; SHR: Spontaneously hypertensive rat; 2K1C: Two-kidney one-clip

In hypertensive adults and children, plasma H₂S levels [240, 246, 247] and expression of CSE were constantly low [240]. In addition, plasma H₂S levels are also reportedly low in patients with portal HTN [248], pulmonary HTN [249], women with preeclampsia [250], and lead-induced HTN [251]. However, although H₂S are generally low in HTN, underlying mechanism(s) of this remain to be fully revealed.

Hydrogen sulfide in neuroinflammation

H₂S is a neuromodulatory and neuroprotective molecule [253]. As it can freely cross the cell membrane, it can regulate various intracellular signaling processes [253]. H₂S can act as an endogenous neuromodulator via enhancing NMDA receptor-mediated responses [254, 255] and by modulating intracellular pH [256] and Ca²⁺ levels in neurons [257], astrocytes [258] and microglia [259]. In addition, H₂S donors have shown beneficial effects on glia-mediated neuroinflammation in various neurodegenerative conditions [260–264] and HTN [108, 265] via their anti-inflammatory [108, 260, 261, 263, 264], anti-oxidant [108, 265] and anti-apoptotic effects [261]. Thus, at physiological levels, H₂S exerts largely beneficial effects on the brain. However, if the brain concentration of H₂S deviates from its physiological range, this can lead to pathology and/or toxicity [229, 274, 276].

Restoration of diminished H₂S levels or the H₂S-producing enzyme activity in the brain is beneficial in various neurodegenerative conditions [264, 266–268]. For example, decreased CBS expression has been reported in the rostral ventrolateral medulla (RVLM) of SHR, and microinjection of NaHS or CBS agonist S-adenosyl-l-methionine into the RVLM decreased BP by decreasing the NADPH oxidase activity. On the contrary, microinjection of CBS inhibitor hydroxylamine hydrochloride increased BP [265]. One study suggested that decreased CBS expression and H₂S production in the RVLM of SHR might be a compensatory mechanism of HTN involving the H₂S-NO cross talk [269], but this remains to be confirmed. In addition, decreased CBS expression and H₂S production was observed in the PVN of high salt-fed hypertensive rats, and microinjection of H₂S donor GYY4137 into the PVN restored CBS expression and H₂S production while reducing the sympathetic activity and BP in these rats [108]. A decreased CBS activity and H₂S production were also reported in astrocytes of stroke-prone SHR compared with the SHR [270]. H₂S has also been shown to protect BBB integrity following experimental cerebral ischemia by suppressing local inflammation and ROS generation [271], or during cardiac arrest by suppressing matrix metalloproteinase-9 and vascular endothelial growth factor expression while increasing Ang-I expression and tight junction expression [272, 273], thus preventing inflammatory factors from crossing the BBB and causing neuroinflammation. Taken together, these findings indicate that disruption of endogenous H₂S generation in the brain may be involved in pathogenesis of HTN. However, there remains a need to understand the precise underlying mechanisms causing this disruption.

Hydrogen sulfide in the gut

Biosynthesis of hydrogen sulfide in the gut

H₂S in the GI tract is produced through enzymatic reaction in host epithelial cells [274] and by the gut microbiota [275]. Enzymatically produced H₂S is primarily derived from sulfur-containing amino acids methionine, homocysteine and cysteine [276] by CBS and CSE, and cysteine aminotransferase in conjunction with mercaptopyruvate sulfurtransferase (3-MST) [277, 278]. Expression of these enzymes differs in tissues. CBS is mainly expressed in the liver, colon and CNS, CSE is mainly found in smooth muscle cells in the cardiovascular system and in the GI tract, and 3-MST is expressed mainly in the brain [279, 280].

Several different genera of colonic bacteria are capable of producing the H₂S. Gut bacteria produce H₂S by enzymatic degradation in the large intestine while utilizing the sulfur-containing amino acids, inorganic sulfate, sulfited additives and sulfated polysaccharides such as sulfomucins [275]. In this way, Enterococci, Enterobacteria and Clostridia including Escherichia coli produce H₂S by cysteine desulfhydrase activity [275]. Metabolism of sulfite by sulfite reductase is another mechanism of H₂S production. Sulfate-reducing bacteria (SRB) are a heterogeneous group of microorganisms which use sulfate as an electron acceptor in the process of dissimilatory sulfate reduction resulting in the production of H₂S [281, 282]. SRB may be divided into four groups based on rRNA sequence analysis: Gram-negative mesophilic SRB (within the delta subdivision of the Proteobacteria: Desulfobulbus, Desulfomicrobium, Desulfomonas, Desulfovibrio, Desulfobacter, Desulfobacterium, Desulfococcus, Desulfomonile, Desulfonema, Desulfosarcina); Gram-positive spore forming SRB (Desulfotomaculum); thermophilic bacterial SRB (Thermodesulfobacterium); and thermophilic archaeal SRB (Archaeoglobus) [283]. For sulfate reduction, SRB need exogenous electron donors, including hydrogen (H₂), methanol, ethanol, acetate, lactate, propionate, butyrate and sugar [284]. A study with germ-free mice showed reduced plasma, cecum and colon H₂S levels compared to the conventional mice, concluding that gut microbiota contributes to circulating and GI H₂S levels. Moreover, absence of the gut microbiota in germ-free mice was associated with reduced CSE activity in several tissues including the cecum, colon, small intestine, kidney, liver, aorta, heart and brain, suggesting that gut microbiota may also influence CSE activity or expression in the host tissues [285].

Multiple biological roles of hydrogen sulfide in the gastrointestinal tract

In the GI tract, H₂S is an important contributor to the maintenance of the mucosal barrier. For example, H₂S can protect against gastric mucosal injury by increasing gastric mucosal blood flow and preventing leukocyte adherence to the endothelium of mesenteric vasculature [279], suggesting vascular and immune targets in the GI tract. H₂S can protect mucosal integrity by inducing bicarbonate secretion, possibly via stimulating the capsaicin-sensitive afferent neurons resulting in NO and prostaglandin E₂ release [286, 287], suggesting effects on the peripheral nervous system within the GI tract. Indeed, H₂S has a neuromodulatory role in the gut by acting on the transient receptor potential vanilloid 1 receptors on primary afferent terminals in the GI tract [288], and causing excitation of the sensory nerves by activating Ca(v)3.2 T-type Ca(2+) channels considered to be protective in rats with colitis [289]. H₂S also acts as an energy source for the GI epithelial cells [290], in addition to its anti-oxidant role and a role in regulation of the cell cycle [291].

4. A potential role for hydrogen sulfide in gut-brain axis in hypertension

Evidence thus far indicates that (i) gut microbiota and neuroinflammation are involved in pathogenesis of HTN; (ii) bidirectional communication exists between gut and brain; and (iii) H₂S levels are reduced in HTN, while studies to date have largely focused on the enzymatic production and activity in circulation and the host tissues. Thus, the contribution of gut-derived H₂S in the pathogenesis of HTN is not known. The following questions remain unanswered: Is there a link between gut microbiota-derived H₂S and HTN? What is the role of H₂S-producing bacteria in HTN and how may the modulation of H₂S-producing bacteria impact HTN? Can we alleviate neuroinflammation by exogenous H₂S administration in HTN? Answers to these questions may aid in the development of dietary or pharmaceutical gut-brain targeted approaches to prevent or control high BP.

Hydrogen sulfide and gut dysbiosis in hypertension

Recently, experimental intracolonic administration of H₂S donors lowered BP, suggesting gut derived-H₂S may contribute to BP control [292, 293]. As mentioned before, gut dysbiosis is present in HTN; however, the evidence of contribution of the H₂S-producing bacteria to the hypertensive phenotype is scarce. Desulfovibrio, the predominant SRB genus, shows higher abundance in patients with HTN [184, 185], while in the SHR the abundance of Desulfovibrio does not correlate with BP [294]. As discussed above, circulating levels of H₂S are reportedly reduced in HTN including the SHR (Table 2). We measured fecal H₂S using the modified methylene blue method [295], and our new preliminary data show that fecal H₂S levels are also reduced in the SHR (Figure 1) [unpublished data]. Thus, fecal H₂S levels could be an important and direct biomarker of H₂S production by the gut bacteria as well as an indicator of circulating H₂S in HTN.

Figure 1.

A) Schematic representation of the currently available methods of hydrogen sulfide (H₂S) detection in biological samples, including colorimetric methods [296–298], gas chromatography [299, 300], high performance liquid chromatography (HPLC) [301–303], headspace-gas chromatography-mass spectrometry (HS-GC-MS) [304], liquid chromatography-mass spectrometry (LC-MS/MS) [305–307], polarographic sensors [308], fluorescent probes [309, 310], enzyme-linked immunosorbent assay (ELISA) [311, 312]. These detection methods have several limitations, including detection sensitivity, expense, sample type etc. [313, 314]. Accurate measurement of biologically active H₂S is challenging since H₂S is a volatile gas and exists in the body in various forms, including free, acid-labile, and sulfane sulfur bound [315]. B) Methylene blue method [295] was used for the measurement of fecal H₂S in spontaneously hypertensive rats (SHR). Fecal H₂S levels were significantly lower in the adult male SHR compared to the age- and sex-matched normotensive Wistar-Kyoto rats (SHR: 0±0.01703 Arbitrary Unit (AU) vs. WKY: 0.094±0.03385 AU, n=5 per strain; *p<0.05 by T-test, mean ± SEM).

Indirect but supporting evidence to our current findings are that plasma levels of sulfur-containing amino acids (taurine, serine, methionine and threonine) are reportedly decreased in patients with essential HTN [316]. This may have resulted from impairment in sulfur amino acid metabolism; however, the underlying mechanism is not yet known. Interestingly, Desulfovibrio is elevated in a small group of hypertensive patients [185] and patients with inflammatory GI disorders such as ulcerative colitis and irritable bowel disease [317–319] and neurodegenerative diseases [320–322]. This increase in SRB Desulfovibrio may be related to inflammation of the gut present in the GI disorders [317–319]. SRB can reduce disulfide bonds thereby causing mucin denaturation and microbial access to the mucus layer [323]. Thus, high non-physiologic levels of H₂S could disrupt the gut barrier function and cause detrimental effects such as colonic pain, gastrointestinal discomfort, inflammatory bowel disease, colonic nociception, and colorectal cancer [324–327]. Further studies employing genomic sequencing of SRB and investigation of fecal and host H₂S levels in rodents and humans with HTN are needed to clarify this matter.

HTN-associated gut dysbiosis is characterized by an imbalance in specific microbial populations and their corresponding metabolites, specifically decreased acetate and butyrate-producing bacteria and increased lactate-producing bacteria [146]. SRB can utilize these organic compounds as electron donors for sulfate reduction [328]. In this way, NaHS can inhibit n-butyrate oxidation by inhibiting short chain acyl-CoA dehydrogenase, which is important in maintaining colonic mucosal integrity [329, 330]. In addition, reduced Desulfovibrio was beneficial in mice with phenylketonuria as it increased cecal levels of acetate, butyrate, and propionate while reducing plasma PIC levels, suggesting improved intestinal barrier function [331]. Moreover, SRB competition for organic compounds such as lactate can stimulate sulphide formation by SRB [332]. Interestingly, both SRB and lactate-producing bacteria are increased in HTN. Consequently, reducing the abundance of SRB may be beneficial for the protection of gut-integrity. However, there remains a need to clarify cause-effect relationships between SRB and HTN.

Supplementation and dietary modification as anti-hypertensive treatment options

Experimental and clinical studies suggest that H₂S donors or dietary supplementation with sulfur-containing amino acids, the precursors of H₂S, may be beneficial for HTN and neuroinflammation. For instance, injection of H₂S donors such as NaHS and STS reduces BP in Ang II-dependent HTN and the SHR [233, 252, 333, 334], while D-cysteine or L-cysteine supplementation via gastric gavage results in decreased BP and reduces kidney damage induced by high-salt diet in the SHR [335]. Moreover, garlic contains a number of active sulfur compounds and has been shown to reduce BP in hypertensive individuals [336, 337], which may be partly mediated via increase in H₂S production [338]. In pre-hypertensive and borderline HTN humans, chronic taurine supplementation, a sulfur-containing amino acid, has been shown to lower BP, which was correlated with an increase in plasma H₂S levels and upregulated vascular H₂S-producing enzymes [339, 340]. Taurine supplementation has also been shown to reduce glial activation and levels of inflammatory mediators while increasing H₂S levels and CBS expression in rats with intracerebral hemorrhage [341, 342]. Methionine is another sulfur-containing essential amino acid. However, the effects of dietary methionine intake on BP remain tenuous. Some studies linked elevated dietary methionine intake with increase in BP in normotensive rats [343–345], whereas it reduced BP in the SHR [346] and DOCA rats [347]. Methionine overloading has also been shown to attenuate development of HTN in female SHR [348]. On the other hand, restriction of sulfur amino acids methionine and cysteine can enhance endogenous CSE activity resulting in increased H₂S production and protection from hepatic ischemia reperfusion injury in mice [349]. Thus, the effects of dietary sulfur-containing amino acids may differ in physiological and pathological conditions and could involve the interaction of various compensatory mechanisms. These studies predominantly focus on the enzymatic production of H₂S by the host, while the gut microbiota can also contribute to circulating H₂S levels. Absence of the gut microbiota in germ-free mice was associated with reduced CSE activity in several tissues including the cecum, colon, small intestine, kidney, liver, aorta, heart and brain, suggesting that gut microbiota may also influence CSE activity and/or expression in host tissues [285]. H₂S production in the gut also depends on bacterial substrates. For instance, sulfur-amino acid cysteine is reportedly more powerful and more reliable in stimulation of H₂S production compared to the inorganic sulfur [350, 351]. Thus, these factors should also be considered while attempting to manage H₂S production in the host.

In addition to the influence of dietary sulfur-containing amino acids on H₂S production, other types of diets can also affect this process. For instance, in the high fat Western diet-fed rat model, L-cysteine-induced aortic vasorelaxation is significantly inhibited [352]. Moreover, diets high in protein can increase SRB abundance and consequently the fecal sulfide levels [353]. Although diet can impact endogenous H₂S production and H₂S-generating bacteria, the dietary requirements and bacterial manipulations needed for restoring the H₂S deficiency in HTN are currently unknown. Thus, the potential contribution of H₂S-producing bacteria in dietary H₂S production in HTN needs to be clarified in future studies.

4. Conclusion and future prospects

With reported neuromodulatory, anti-inflammatory and anti-hypertensive properties, both host- and gut bacterial-derived H₂S may be affected in HTN. Low plasma H₂S levels are reported in human hypertensives and rodent models of HTN, and our new preliminary data show that this is also reflected in reduced fecal H₂S in the SHR. However, contribution of gut dysbiosis and gut microbiota-derived H₂S to the pathology of HTN remains unknown. Future studies should aim to provide detailed understanding of how the levels of H₂S may be modified with dietary manipulations including application of exogenous H₂S donors in order to manipulate bacterial and host production of H₂S as a potential therapy for HTN.

Abbreviations:

Ang

Angiotensin

H₂S

Hydrogen sulfide

PNA

Parasympathetic nerve activity

SNA

Sympathetic nerve activity

ROS

Reactive oxygen species