Role of vasoactive substances on endometrial and ovarian function

Abstract

In this review, it is proposed that the vasoactive agents endothelin (ET), nitric oxide (NO)/NO synthase (NOS) and carbon monoxide(CO)/heme oxygenase(HO) act directly on human endometrial functions and on ovarian functions in the normal menstrual cycle and in implantation periods. These vasoactive substances are likely to be important autocrine/paracrine factors that regulate a variety of physiological and pathological processes. The main actions of these agents are differentiation and implantation in the endometrial functions, and follicular growth, luteinization and atresia in the ovarian functions, in the tight connection between endometrial and ovarian systems during normal menstrual periods and during implantation (Reprod Med Biol 2007; 6: 157–164)

INTRODUCTION

THE AMOUNT OF hormones secreted from the corpus luteum in the luteal phase has been shown to be largely influenced by the process of follicle maturation and early luteinization. In addition, full maturation of the secretory endometrium is affected by various factors, especially sex steroids secreted from the corpus luteum and some physiologically active substances. Recently, some vasoactive substances derived from vascular endotherial cells have been implicated in a variety of physiological and pathological processes. Therefore, vasoactive agents may act directly on endometrial blood vessels during the secretory phase of a normal menstrual cycle and implantation.

Endothelin (ET) was originally isolated from the supernatant of cultured porcine vascular endothelial cells ¹ and is a potent and long‐lasting vasoconstricting peptide made up of 21 amino acids. The human ET family is composed of three isopeptides, ET‐1, ET‐2 and ET‐3. ² Two specific receptors for ET (ETA, ETB) ³ , ⁴ and messenger RNA (mRNA) for prepro‐ET‐1, the precursor of ET‐1, are widely distributed in a variety of human tissues, ⁵ , ⁶ suggesting that ET‐1 has multiple physiological functions not only in the cardiovascular system but also in the non‐vascular system. Many reports have suggested that ET‐1 has a functional role in the female reproductive system, ⁷ particularly in the human endometrium. ⁸ However, to the best of our knowledge, very few studies examining the relationship between the ET‐1 secretion and nitric oxide (NO)/NO synthase (NOS) system in the endometrium have been reported to date.

Nitric oxide, a highly reactive free radical gas, has recently been implicated in a variety of physiological and pathological processes. Nitric oxide is synthesized by conversion of l‐arginine to l‐citrulline by a family of enzymes known as NOS, ⁹ which includes: (i) a calcium‐independent, inducible NOS (iNOS), (ii) a constitutive, calcium‐dependent endothelial NOS (eNOS); ¹⁰ and (iii) a constitutive, neural NOS (nNOS). Nitric oxide is recognized as playing an important role almost everywhere in the human body, including the central nervous system, immune system, platelets, ¹⁰ kidney ¹¹ and gastrointestinal system. ¹² In contrast, carbon monoxide (CO), as well as NO, is a novel gaseous chemical messenger that plays key roles in cell function and cell–cell communication in many organ systems. ¹³ , ¹⁴ The formation of endogenous CO occurs during the metabolism of intracellular heme to biliverdin and iron, which is catalyzed by heme oxygenase (HO). ¹³ Biliverdin is subsequently converted to bilirubin by biliverdin reductase. Heme oxygenase mainly consists of three distinct isozymes, HO‐1, HO‐2 and HO‐3. HO‐1 is induced not only by its substrate heme, but also by various kinds of oxidative stress. ¹⁵ HO‐1 has been shown to be an oxygen‐related protein 33. ¹⁶ This isoform is highly expressed in stimulated liver ¹³ and the spleen, where it is responsible for the destruction of heme from red blood cells. ¹⁷ In contrast, HO‐2 is constitutively expressed and is distributed in unstimulated liver, ¹³ the testis ¹⁸ and the brain at the highest concentration. ¹⁹ The two forms are products of two different genes and differ in their tissue expression. HO‐3 has very low activity, and its physiological function probably involves heme binding. ²⁰ The recognition of NO as an important gaseous messenger of cell signaling has led to a reappraisal of the likely role of endogenous CO. Nitric oxide and CO signaling processes have been demonstrated in the central nervous system, ²¹ and both NO and CO induce vasodilatation via the activation of soluble guanylate cyclase, and ensure relaxation of vascular smooth muscle in the cardiovascular system. ²²

This review was undertaken to investigate the ET system, the NO/NOS system and the CO/HO system on the differentiation induced by sex steroid hormones in human endometrial cells in normal menstrual periods and human decidual cells in implantation, and on follicular development or luteinization in human or porcine granulosa cells, using mainly molecular biological methods.

EFFECT OF ENDOTHELIN ON THE HUMAN ENDOMETRIUM IN A NORMAL MENSTRUAL CYCLE

IT HAS BEEN demonstrated that mRNA for the prepro‐ET‐1 and the known receptor subtype (ETA and ETB) are present in human endometrium at different stages of the menstrual cycle obtained by hysterectomy. ²³ , ²⁴ , ²⁵ Northern blot analysis revealed expression of ET‐1 mRNA in human endometrium during a normal menstrual cycle. The concentration of ET‐1 mRNA in endometrial tissue was greater during the menstrual and proliferative phases than during the ovulatory and secretory phases. Immunoreactive ET‐1 was secreted into the medium of isolated endometrial stromal cells. Estradiol and progesterone significantly attenuated ET‐1 release in endometrial stromal cells cultured for 6 days. On the contrary, prolactin (PRL) release increased in a time‐dependant manner in the same experiments, and there was a significant negative correlation between the release of PRL and ET‐1 release. ²⁶ ETA and ETB mRNA were also present in endometrial tissue of a normal cycle. The concentration of ETA receptor mRNA was greater in the proliferative phase than in the secretory phase, whereas expression of ETB mRNA increased in the menstrual phase. ²³ ET‐1 significantly increased extracellular accumulation of cyclic AMP (cAMP), intracellular generation of inositol phosphates and significantly enhanced DNA synthesis in cultured endometrial stromal cells from the proliferative phase. ²³ These results showed that human endometrial cells synthesized and released ET‐1, and contained ETA and ETB receptors that were functionally coupled to phosphoinositide breakdown and to adenylate cyclase with the increase of cAMP by ET‐1 stimulation. These findings suggest that ET‐1 may have a potential autocrine and/or paracrine function in human endometrial stromal cells.

The physiological functions of ET‐1 in human endometrium have been suggested as follows: ¹ ET‐1 acts during menstruation to restrict blood loss to promote cessation of endometrial bleeding; ² , ²⁴ ET‐1 produced in the endometrium reaches the myometrium and stimulates uterine contractions; ³ , ²⁷ ET‐1 may serve to promote re‐epithelialization; ²⁸ and ⁴ ET‐1 produced by the glandular epithelium may be involved in blastocyst apposition/implantation or trophectoderm replication. ²⁴

EFFECT OF ENDOTHELIN ON THE HUMAN DECIDUA IN IMPLANTATION PERIODS

IT HAS BEEN reported that a novel vasoconstrictor, ET‐1, was synthesized and released from human cultured decidual cells in early pregnancy during a whole cultured period. ⁷ Reverse‐phase, high‐pressure liquid chromatography of the conditioned media from the decidual cells revealed a major peak of ET‐1 coeluting with standard ET‐1. Phorbor myristate acetate, a protein kinase C activator, dose‐dependently increased the release of ET‐1 from the decidual cells, while a protein kinase C inhibitor, H7, significantly attenuated the stimulatory effect of phorbor myristate acetate on ET‐1 release. Northern blot analysis demonstrated the expression of mRNA for prepro‐ET‐1 in the decidual tissue. The human decidual tissue contained a non‐interacting, single class of binding sites, demonstrating higher affinity for ET‐1 and ET‐2 than ET‐3. This would be most consistent with the ETA receptor subtype. An ET‐1‐induced, dose‐dependent accumulation of total inositol phosphates was also observed in human decidual cells prelabeled with myo‐[³H] inositol. These results demonstrated that human decidual cells in early pregnancy can synthesize and release ET‐1. These cells also possess specific functional receptors for ET‐1 that are coupled to phosphoinositide hydrolysis. Thus, this report suggests a possible role for ET‐1 in autocrine and/or paracrine function in human decidual cells. ⁷

Recently, it has been shown that ET‐1 was detectable in the amniotic fluid of term pregnancy, ²⁹ and that prepro‐ET mRNA was expressed in the avascular human amnion at term. ³⁰ Taken together, these observations suggest that ET‐1 present in human amniotic fluid may be derived primary from decidual cells.

EFFECT OF THE NO/NOS SYSTEM ON HUMAN ENDOMETRIUM IN A NORMAL MENSTRUAL CYCLE AND DECIDUA IN IMPLANTATION PERIODS

THE EFFECT OF NO/NOS on the differentiation of human endometrial cells has previously been reported. Nested reverse transcriptase‐polymerase chain reaction (RT‐PCR) produced one intense band of human iNOS mRNA and human eNOS mRNA from total RNA extracted from the endometrium. However, northern blot analysis revealed no expression of iNOS mRNA in human endometrium of a normal menstrual cycle. ³¹ eNOS‐like immunoreactivity was detected in endothelial cells lining blood vessels in 73% (11/15) of the endometrium in the secretory phase versus 37% (7/19) of the endometrium in the proliferative phase in an immunocytochemical study. ³² There has been a previous study in humans that revealed protein and mRNA for eNOS detected in endometrial glandular epithelium and stroma by in situ hybridization and immunocytochemistry. ³³ In contrast, iNOS‐like immunoreactivity was not detected in the human endometrium, ³² while the expression of iNOS mRNA has only been detected in the epithelial glands of a menstrual endometrium using solution hybridization/ribonuclease protection assay. ³⁴ Furthermore, endometrial stromal cells were cultured in the absence or the presence of cytokines, such as interleukin‐1β (IL‐1β) and interferon‐γ (INF‐γ), in the endometrium. Nested RT‐PCR detected iNOS mRNA in stromal cells cultured without cytokines. Northern blot analysis failed to detect iNOS mRNA in stromal cells cultured for 9 h in the absence of cytokines or in the presence of IL‐1β or INF‐γ alone, but could detect iNOS mRNA cultured with a combination of IL‐1β and INF‐γ. Western blot analysis demonstrated iNOS protein in stromal cells cultured for 12 h with combined IL‐1β and INF‐γ. ³¹ These results raise the possibility that NO locally synthesized by iNOS may be involved in the control of endometrial functions.

In the next step, the expression and distribution pattern of iNOS in non‐pregnant and early pregnant human endometrium were studied using RT‐PCR and immunohistochemistry. Northern blot analysis revealed the expression of iNOS mRNA in human decidua and chorionic villi in the first trimester, but not in human endometrium at all stages of the menstrual cycle. ³⁵ Immunohistochemical staining of the secretory endometrium using an antihuman iNOS polyclonal antibody revealed labeling specifically concentrated in glandular epithelial cells. Staining was absent in stromal cells. ³⁵ However, iNOS staining was positive in decidualized stromal cells in tissues obtained in the first trimester of pregnancy. Furthermore, extensive staining was observed in both syncytiotrophoblastic and cytotrophoblastic cells. ³⁵ The finding of a large amount of iNOS mRNA at the feto–maternal interface throughout the first trimester of pregnancy suggests that iNOS may play an important role in the maintenance of pregnancy.

Studies have also revealed that the NO/NOS system in human endometrium is involved in the regulation of ET‐1 release via IL‐1β. It can be inferred that NO and ET‐1 control the functions of the endometrium in close association with IL‐1β. ³⁶

EFFECT OF THE CO/HO SYSTEM ON HUMAN ENDOMETRIUM IN A NORMAL MENSTRUAL CYCLE AND PLACENTAL VILLI IN IMPLANTATION PERIODS

THE PRESENCE OF HO‐1 and HO‐2 in human endometrium was investigated at various stages of the menstrual cycle using RT‐PCR, western blotting and immunohistochemistry. The RT‐PCR detected mRNA for HO‐1 and HO‐2 in human endometrium at all stages of the menstrual cycle. Western blotting also revealed the expression of the two distinct HO proteins throughout the menstrual cycle. HO‐1 was constitutively expressed, whereas HO‐2 expression was apparently greater in the secretory phase than in the menstrual and proliferative phases. ³⁷ Immunohistochemistry revealed that the distribution of the two HO isoforms had distinct topographic patterns: HO‐1 was observed in endometrial epithelial cells and macrophages, whereas HO‐2 was found in endothelial cells and smooth muscle cells of blood vessels in the endometrium. ³⁷ It has been reported that endometrial macrophages increased in number late in the secretory phase and in early pregnancy. ³⁸ Cytokines produced by macrophages are reported to activate iNOS in human endometrium ³¹ and to induce HO‐1. ³⁹ The synergetic biological effects of HO‐1 and iNOS induced by macrophages are likely to contribute to some physiological functions of the human endometrium. The detection of mRNA and protein for HO‐1 and HO‐2 in normal human endometrium suggests that the CO/HO system may play a role in the local control of endometrial function. ³⁷

In the second step, the effect of CO/HO on implantation was investigated. It has been elucidated that RT‐PCR detected mRNA for HO‐1 and HO‐2 in human villi in the first trimester of pregnancy, and that western blotting also revealed the expression of the two distinct HO proteins in the first trimester of pregnancy. ⁴⁰ HO‐1 was continuously expressed throughout gestation, while HO‐2 expression was apparently weaker in the first trimester than at term. Immunohistochemistry revealed that the distribution of the two HO isoforms had distinct topographic patterns: HO‐1 was observed in villous trophoblastic cells, while HO‐2 was found in endothelial cells and smooth muscle cells of blood vessels of placental villi. ⁴⁰ In contrast, another report has shown that HO‐1 and HO‐2 were exclusively expressed in the trophoblast in rat placenta and that mRNA for the two HO isoforms gradually decreased during gestation. ⁴¹ These results may provide a microtopographic basis for elucidating the mechanism of CO‐mediated vasodilatation, and it is suggested that the CO/HO system may be involved in the control of placental vascular function and may protect the syncytiotrophoblast and endothelium against oxidative injury. ⁴⁰

EFFECT OF ET/ET RECEPTOR ON OVARIAN FUNCTION

STUDIES HAVE EXAMINED whether the novel vasoconstrictor ET‐1 is synthesized by and released from porcine granulosa cells (PGC), and whether ET‐1 acts directly on granulosa cells in an autocrine/paracrine fashion. ET‐1 was released from cultured PGC during a whole cultured period. ⁴² The dilution curve constructed using the conditioned medium of the cultured PGC was parallel to a standard curve of ET‐1 in radioimmunoassay. Northern blot analysis showed the expression of mRNA for prepro‐ET‐1 in PGC. ⁴² Scatchard analysis of a binding study of cultured PGC using ¹²⁵I‐labeled ET‐1 indicated the presence of a single class of high‐affinity binding sites with almost equal affinity for ET‐1 and ET‐3, that is ETB receptor. ⁴³ Dose‐dependent ET‐1 induced rapid and transient increases in intracellular Ca²⁺ in fura‐2‐labeled cells, and ET‐1 also dose‐dependently stimulated total inositol phosphates in cells prelabeled with myo‐[³H]inositol. ⁴³ In addition, ET‐1 dose‐dependently stimulated cell growth and DNA synthesis in PGC. ET‐1 inhibited the follicle stimulating hormone (FSH)‐stimulated and human chorionic gonadotropin‐stimulated accumulation of progesterone in 24‐h PGC incubation. ⁴² These results suggest that ET‐1 is a potent inhibitor of differentiation, and possibly assists follicular development by inhibiting luteinization in PGC. Dose‐dependent inhibition of luteinizing hormone‐stimulated or FSH‐stimulated cAMP accumulation by ET‐1 in PGC has been reported. ⁴⁴ , ⁴⁵ The presence of ET‐1 in the follicular fluid of porcine ovaries has been demonstrated, and it has been elucidated that the concentration of ET‐1 in large follicles (>5 mm) was higher than that in small (<3 mm) and medium (3–5 mm) follicles. ⁴² These findings clearly demonstrate the presence of an ETB receptor in PGC, coupled with phosphoinositide hydrolysis and intracellular Ca²⁺ mobilization, and that ET‐1 produced by PGC may function as an autocrine/paracrine growth factor and modulator of steroidogenesis in ovarian granulosa cells.

The effect of ET‐1 on luteinized human granulosa cells (L‐HGC) obtained from patients undergoing in vitro fertilization has been reported. ET‐1 increased intracellular Ca²⁺ in L‐HGC in a dose‐dependent manner. Messenger RNAs for two known receptor subtypes (ETA and ETB) were also present in L‐HGC; however, the expression of ETA receptor mRNA was much greater than that of the ETB receptors. ⁴⁶ ET‐1 stimulated cell proliferation in L‐HGC in a dose‐dependent manner and these stimulatory effects were completely blocked by BQ‐123, an ETA receptor antagonist. ET‐1, ET‐3 and IRL‐1620, a selective ETB receptor agonist, attenuated basal progesterone secretion in L‐HGC. These results have revealed that ETA receptors predominantly exist in L‐HGCs and that ET‐1 may stimulate cell proliferation of L‐HGC by increasing intracellular Ca²⁺ via ETA receptors. ⁴⁶ In addition, it should be noted that large amounts of ET‐1 have been identified in follicular fluid obtained from mature ovarian follicles. ⁴⁷

EFFECT OF THE NO/NOS SYSTEM ON OVARIAN FUNCTION

RECENTLY, IT HAS been shown that NO may regulate some ovarian functions. The ovarian NO/NOS system seems to be necessary for pre‐ovulatory follicular expansion because of a good correlation between follicular nitrate/nitrite concentration and follicular fluid volume, ⁴⁸ and necessary for follicle rupture during ovulation because NOS inhibitors suppress human chorionic gonodotropin‐induced ovulation. ⁴⁹ Moreover, other reports have shown that NO suppresses apoptosis in cultured pre‐ovulatory ovarian follicles, ⁵⁰ and that NO inhibits steroidogenesis in human granulosa luteal cells as an autocrine regulator. ⁵¹ The effect of NO/NOS on the follicular maturation and luteinization induced by gonadotropin (Gn) has been reported in PGC derived from medium‐sized (3–5 mm) ovarian follicles. ⁵² An NO donor, NOC18, significantly suppressed the estradiol (E2) release from basal and Gn‐stimulated PGC in a 2‐h culture. In addition, NOC18 significantly inhibited the aromatase activity of basal and Gn‐stimulated PGC as measured using a modified tritiated water method. However, the cGMP analog, 8‐bromo‐cGMP, had no significant effect on the accumulation of E2 and progesterone (P) during a 24‐h culture. A NOS inhibitor, N^G‐monomethyl‐L‐arginine (LNMMA), significantly stimulated the basal E2 release and dose‐dependently enhanced the E2 and P release from FSH‐stimulated PGC in a 24‐h culture. However, N^G‐monomethyl‐d‐arginine, which does not inhibit NOS, did not enhance the release of E2 and P under the same experimental conditions. In addition, LNMMA significantly suppressed nitrite concentrations in the media as measured using chemiluminescence. These results demonstrated that NO inhibits E2 secretion independent of cGMP by inhibiting P450 aromatase activity in moderately mature PGC. ⁵²

Moreover, an additional report investigated the NO/NOS system on steroidogenesis in PGC derived from small follicles (<3 mm) (S‐PGC) in comparison with those derived from medium follicles (3–5 mm) (M‐PGC). ⁵³ NOC18 significantly suppressed basal and Gn‐stimulated E2 release from both S‐PGC and M‐PGC in a 2‐h culture. NOC18 significantly decreased basal and Gn‐stimulated P release from S‐PGC, but not from M‐PGC. In addition, NOC18 significantly inhibited aromatase activity in S‐PGC. LNMMA had a significantly stimulatory effect on the basal release of E2 and P from M‐PGC; however, it had no significant effect on basal steroidogenesis in S‐PGC in a 24‐h culture. In the presence of Gn, LNMMA significantly stimulated the release of E2 and P from both S‐PGC and M‐PGC, and this stimulatory effect was weaker in S‐PGC than in M‐PGC. These results demonstrate that NO inhibits E2 secretion by directly inhibiting the aromatase activity in S‐PGC, as in M‐PGC. It has been shown that the NO system suppresses the differentiation of S‐PGC; however, the extent of suppression decreased with the progression of follicular growth. Furthermore, the activity of NOS in S‐PGC was weaker than that in M‐PGC. It has been strongly suggested that the NO/NOS system in PGC regulates steroidogenesis differently during the different phases of follicular development. ⁵³

Studies have also reported that both eNOS and iNOS are detected in ovarian cells. Van Voorhis et al. ⁵⁴ demonstrated that iNOS mRNA expression was elevated at a time when steroidogenesis was virtually absent in the rat ovary, and that in contrast to iNOS, eNOS mRNA levels increased after Gn stimulation and peaked in ovaries containing ovulatory follicles.

EFFECT OF THE CO/HO SYSTEM ON OVARIAN FUNCTION

AS CO AND NO have similar cellular functions in several organ systems, the CO/HO system may also be involved in the follicular development of the ovaries. The possibility that HO might stimulate steroidogenesis in rat ovaries has been recently reported. ⁵⁵ In addition, the expression and possible role of HO in PGC has been elucidated. Harada et al. demonstrated that two HO isozymes, inducible HO‐1 and constitutive HO‐2, are expressed in PGC throughout follicular development; HO‐1 protein was mainly expressed in PGC in atretic follicles, whereas HO‐2 protein was constitutively observed in PGC of the healthy follicle. ⁵⁶ Furthermore, these researchers reported the detection of both HO isozymes in freshly isolated PCG during follicular development using RT‐PCR and western blotting, and demonstrated the presence of both isozymes in two subpopulations of PGC (tightly bound and weakly associated PGC) that were freshly isolated and cultured using flow cytometric analysis. The effect of HO on apoptosis of granulosa cells was also studied using flow cytometry to detect subdiploid DNA fluorescence, and the effect on the expression of Fas ligand was investigated using a quantitative analysis of western blotting and flow cytometry. In tightly bound PGC, the mean proportion of apoptotic cells treated with 1 µmol/L hemin (a HO substrate) was approximately 1.7‐fold greater than that in untreated controls, and zinc protoporphyrin IX (ZnPP IX; a HO inhibitor) completely inhibited the increase in apoptosis induced by hemin in 24‐h culture. Conversely, in weakly associated PGC, the proportion of apoptotic cells was not altered by hemin. ⁵⁶

The Fas/Fas ligand system has been identified as a mediator of granulosa cell apoptosis in the rat ovary, ⁵⁷ , ⁵⁸ and treatment with monoclonal antihuman‐Fas antibody induces apoptosis in cultured human granulosa/luteal cells. ⁵⁹ Fas‐mediated cell‐death pathways are considered to be central to the induction of follicular atresia. ⁵⁹ , ⁶⁰ It has been shown that the quantity of Fas ligand protein increased in a dose‐dependent manner in tightly bound PGC treated with hemin compared with controls, and the hemin‐induced increase in Fas ligand protein was inhibited by ZnPP IX. ⁵⁶

Thus, inducible HO‐1 and constitutive HO‐2 were identified in PGC throughout follicular development, and it has been concluded that products of reactions catalyzed by HOs, namely CO, bilirubin and ferritin, are likely to be important autocrine/paracrine factors that regulate apoptosis in PGC.

CONCLUSION

THE STUDIES REVIEWED here reveal that locally synthesized ET and NO could be involved in the control of endometrial functions, and that the ET and/or NO/NOS systems play an important role in the maintenance of pregnancy. Furthermore, it is suggested that the CO/HO system possibly plays a role in the local control of endometrial function in normal menstrual periods. CO/HO may also be involved in the control of placental vascular function, which protects the syncytiotrophoblast and endothelium against oxidative injury.

In ovarian functions, the results outlined demonstrate that ET produced locally may function as an autocrine/paracrine growth factor and modulator of steroidogenesis in ovarian granulosa cells. Furthermore, it has been elucidated that NO inhibits E2 secretion independent of cGMP by inhibiting P450 aromatase activity in moderately mature PGC, and it is highly likely that the NO/NOS system in PGC regulates steroidogenesis differently during different phases of follicular development. In contrast, the products of reactions catalyzed by the CO/HO system are likely to be important autocrine/paracrine factors that regulate apoptosis in PGC.

In conclusion, some vasoactive substances, ET/ET receptors, NO/NOS and CO/HO, are likely to be important autocrine/paracrine factors that regulate a variety of physiological and pathological processes, including follicular growth, luteinization and atresia in the ovarian functions and proliferation, differentiation and implantation in the endometrial functions, in the tight connection between ovarian and endometrial systems during normal menstrual periods and during implantation (Fig. 1).

Figure 1.

Effect of vasoactive substances on ovarian and endometrial functions. CO, carbon monoxide; ET, endothelin; ETR, endothelin receptor; HO, heme oxygenase; NO, nitric oxide; NOS, nitric oxide synthase; eNOS, endothelial NOS; iNOS, inducible NOS.