| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Molecular and Integrative Physiology and College of Medicine, University of Illinois at Urbana-Champaign, Urbana, Illinois 61801
Correspondence: Address all correspondence and requests for reprints to: Dr. O. David Sherwood, Department of Molecular and Integrative Physiology, University of Illinois at Urbana-Champaign, 524 Burrill Hall, 407 South Goodwin Avenue, Urbana, Illinois 61801. E-mail: od-sherw{at}uiuc.edu
| Abstract |
|---|
|
|
|---|
| I. Introduction |
|---|
|
|
|---|
Research on relaxin has waxed and waned (Fig. 1
). After the economic depression and world conflict in the 1930s and 1940s, there was a surge of research interest on relaxin. From the late 1940s through the early 1960s, impure preparations of relaxin were reported to have numerous effects on the reproductive tract in nonpregnant mammals. The pioneering discoveries that relaxin inhibits spontaneous uterine contractility in estrogen-primed guinea pigs (3), promotes cervical softening in estrogen-primed cattle (4), and promotes elongation of the interpubic ligament in estrogen-primed mice (5) predicted physiological roles that were later established for endogenous relaxin during pregnancy. During the late 1950s and early 1960s, the Warner-Chilcott Laboratories provided an impure preparation of porcine relaxin (Releasin) and supported studies in humans that examined the use of relaxin as a therapeutic agent for three clinical problems (2). It was postulated that, by increasing skin elasticity, relaxin would have beneficial effects in patients with progressive systemic sclerosis. It was also postulated that, by inhibiting uterine contractility, relaxin would prevent premature labor, and, by softening the cervix, the hormone would reduce the duration of labor. Clinical efforts with impure porcine relaxin were not sustained beyond the mid-1960s for several reasons, including lack of consistent effectiveness, safety problems (6), and both the time and expense associated with meeting the new and more stringent regulatory requirements of the United States Food and Drug Administration (2, 7). There was nearly no interest in relaxin for about 10 yr. Then, a marked increase in relaxin research occurred between the mid-1970s and 1980s when improved methods for the isolation and characterization of proteins, as well as recombinant DNA technology, made possible rigorous studies of the chemistry and physiology of the hormone. Nevertheless, publications on relaxin plateaued at a moderate level of approximately 40 publications/yr throughout the last 20 yr.
|
Having said that, it is this reviewers view that several discoveries since the late 1980s have provided exciting new avenues for investigating the actions of relaxin that will increase research on the hormone for the foreseeable future. First, it is now established that relaxins actions are vital during pregnancy in the rat, mouse, and pig (2, 10). Findings in these three species encourage the view that relaxin has important effects in other species. Second, the recent discovery of an additional form of relaxin that is expressed in highest levels in the brain (11) provides inferences that relaxin may act locally as a neurotransmitter within the brain. Third, the recent discovery of the relaxin receptors (12, 13, 14) enables rigorous molecular approaches toward the identification of not only the target tissues for relaxin but also the molecular mechanism(s) whereby relaxin brings about its effects. Finally, the discovery of relaxins actions on nonreproductive tissues such as the brain (15, 16, 17), kidney (18, 19), and heart (20, 21) in both males and females have opened new avenues of investigation not only into relaxins physiological roles but also its therapeutic potential.
B. Scope of review
Before describing the actions of relaxin, the Introduction provides brief descriptions of the structures of relaxin and its receptor, as well as the source and secretion of relaxin in rats, mice, pigs, and humans. Section II describes in physiologically sequential order the actions of relaxin during five female reproductive processes in rats, mice, and pigs. These species offer experimental advantages. In rats and pigs, the corpora lutea contain sufficient relaxin in late pregnancy to enable its isolation in large enough quantities to use for physiological studies. Moreover, the ovaries can be removed to examine the physiological roles of circulating relaxin during pregnancy in both species. Mice have the advantage of using gene targeting to create relaxin-deficient and/or relaxin receptor-deficient animals. Whereas roles for relaxin during follicular growth, ovulation, and implantation have been postulated, the reader is told at the outset that the only well-documented and, in some cases, vital physiological roles of relaxin are those that occur after implantation in pregnant rats, mice, and pigs. Section III describes recently discovered and surprising actions of relaxin on male reproductive processes. Several novel therapeutic actions of relaxin on nonreproductive tissues in male and female rodents have been reported recently. These poorly understood actions, which are described in Section IV, may have clinical potential in humans and other species. Section V contains the conclusions and a view of future directions for relaxin research. For additional information on the history, chemistry, and physiology of relaxin, the reader is referred to other reviews (2, 15, 16, 22, 23, 24, 25) and conference proceedings (26, 27).
C. Structure of rat, mouse, pig, and human relaxin
Relaxin belongs to the insulin/relaxin superfamily of structurally related hormones that in the human includes insulin; IGF-I (28); IGF-II (29); relaxin 1 (30); relaxin 2 (31); relaxin 3 (11); Leydig cell insulin-like peptide (INSL3), which is also designated relaxin-like factor (RLF) (32); early placenta insulin-like peptide (INSL4) (33); INSL5 (34); and INSL6, which is also designated relaxin/insulin factor 1 (35, 36). Whereas the overall sequence identity among family members is low, they are all first synthesized as a prohormone that is comprised of a signal sequence and a B-C-A domain configuration. Within the B and A domains, there are highly conserved cysteine residues that link the A and B domains by two interdomain disulfide bonds and form an A chain intradomain disulfide bond (see the structure of rat relaxin in Fig. 2A
). In several members of the family, including relaxin, INSL3/RLF, and insulin, the C domain peptide is removed during processing of the mature peptide.
|
Recently, human 3 (H3) relaxin and its nearly identical orthologs mouse 3 (M3) relaxin, rat 3 (R3) relaxin, and pig 3 (P3) relaxin were discovered (11, 42, 43). The H3 gene is localized on chromosome 19 at 19p13.3 in close proximity to the RLF gene (11). Relaxin 3 gene is expressed in greatest levels in the ventromedial dorsal tegmental nucleus in the rat and mouse brain, where it has been postulated to act locally as a neuropeptide (13, 42). Synthetic H3 relaxin displays relaxin bioactivity that is about two orders of magnitude lower than that of recombinant H2 (rH2) relaxin in both the human monocyte cell line THP-1 (11) and human 293T cells transfected with H2 relaxin receptors (50). There is evidence that relaxin 3 protein is produced by the brain. Small quantities of P3 relaxin (1015 ng/kg) were isolated from pig brains (13). However, it remains to be determined whether relaxin 3 is of physiological significance in any species. With the exception of the four recently discovered relaxin 3 gene orthologs, which share 90100% amino acid sequence homology (Fig. 2
), there has been limited evolutionary conservation of relaxin among the more than 25 species in which the sequence is known (2, 11, 23). Although rats and mice are in the same taxonomic order (Rodentia), the extent of amino acid sequence identity between R1 relaxin and M1 relaxin is only 75%, and other comparisons of relaxins structure among the four species shown in Fig. 2B
range from 3748%.
Twelve of the amino acids in relaxin are invariant or highly conserved among species. The six cysteine residues involved in disulfide bond formation are invariant (2, 11, 13, 23, 42, 43, 47) with the exception of M1 relaxin, in which an additional tyrosine residue is found penultimate to the C terminus of the A chain (38). The amino acid motif Arg-X-X-X-Arg-X-X-Ile located in the middle of the B chain is required for relaxin bioactivity (2, 51). The arginines in positions B16 and B20 are invariant. Moreover, replacement of isoleucine at B23 is rare and occurs only with large hydrophobic residues such as valine in pig relaxin (40). Finally, three glycines that provide flexibility around cysteine residues are invariant among species.
Because of the structural similarities not only of INSL3/RLF and relaxin but also their receptors, a brief description of the structure of INSL3/RLF and its major physiological role is warranted. INSL3 is also designated RLF because it contains a portion of the putative relaxin receptor-binding region (Arg-X-X-X-Arg), which is offset four amino acid residues toward the C terminus relative to its position in relaxin (52). However, the binding of INSL3/RLF to its receptor appears to require the five-amino acid motif, Gly-Gly-Pro-Arg-Trp, located on the C-terminal end of the B chain (53). INSL3/RLF is produced by testicular Leydig cells and causes descent of the testis during embryonic development (54).
D. Structure of relaxin receptors and signal transduction pathways
For many years, discovery of relaxin receptors proved to be elusive. In 2001 a major breakthrough occurred. This breakthrough was a consequence of an ongoing investigation of a family of leucine-rich guanine nucleotide-binding (G protein)-coupled receptors designated LGRs, which include FSH, LH, and TSH receptors. By using genomics and astute observations of phenotypic expression reported in knockout mice, Hsueh and co-workers and others discovered that two orphan LGR receptors, designated LGR7 and LGR8, are receptors for relaxin (12) and INSL3/RLF (12, 55, 56), respectively (Fig. 3
). LGR7 and LGR8, which are 757 (12) and 737 (55) amino acids in length, respectively, share about 60% amino acid sequence identity and contain 10 leucine-rich repeats in their extracellular domain. Through their activation of LGR7 and LGR8, both relaxin and INSL3/RLF elicit bioactivity, at least in part, through the stimulation of the Gs-cAMP-protein kinase A (PKA)-dependent signaling pathway (12, 56). Available evidence indicates that INSL3/RLF binds to and activates the LGR8 receptor but not the LGR7 receptor (50, 56). Relaxin binds with high affinity and activates the LGR7 receptor (12, 50, 56). Relaxin, however, may not act solely through LGR7. Whereas relaxin binds with greatest affinity to cells containing the LGR7 receptor, it also binds with low affinity to cells that contain only LGR8 receptors (50). Moreover, rH2 relaxin and P1 relaxin induce cAMP production in cells that contain only LGR8 receptors (12, 50, 56). It remains to be determined whether circulating relaxin elicits any biological actions in vivo through LGR8. LGR7 transcripts have been identified in reproductive tissues, as well as nonreproductive tissues such as the brain, kidney, heart, and lung, where actions of relaxin have been reported (Table 1
). A splice variant of LGR7 that is attributable to deletion of an exon located C-terminal to the initial leucine-rich repeat in the extracellular domain of the receptor has been identified (12). Neither the function, if any, nor the tissue distribution of the truncated LGR7 has been reported.
|
|
The intracellular signaling pathways initiated by relaxin are not well understood for any target cell. There is accumulating evidence that relaxin initiates its effects through multiple pathways. Consistent with activation through GPCR activation of Gs, relaxin was reported to increase cAMP and activate PKA in several cells and tissues (2, 22, 63, 64, 65, 66, 67). A recent study with a human monocyte cell line THP-1 provided evidence that relaxin-stimulated cAMP accumulation is biphasic and that activation of phosphoinositide 3-kinase is required for the second wave of cAMP (68). There are also reports supporting the view that relaxin uses a receptor tyrosine kinase-signaling pathway in primary human uterine cells and THP-1 cells (65, 69, 70); the MAPK pathway in human endometrial stromal cells, THP-1 cells and human coronary artery smooth muscle cells (71); and ERK-1/2 in human umbilical vein endothelial cells and epithelial (HeLa) cells (72). Bani and co-workers obtained evidence that relaxin also signals through increased nitric oxide generation and subsequent production of cGMP with studies that used bovine vascular smooth muscle cells (73), rat coronary endothelial cells (74, 75), human basophils (76), human neutrophils (77), human breast cancer cells (78), guinea pig heart (79), mouse uterus (80), and mouse small intestine (81). The significance of distinct signal pathways is presently unknown. It was postulated that they may enable relaxin to be a pleiotropic hormone with a broad range of biological effects on numerous organs and tissues (68).
E. Source and secretion of rat, mouse, pig, and human relaxin
The corpora lutea are the source of circulating relaxin and progesterone throughout pregnancy in the rat, mouse, and pig. Whereas the corpora lutea are also the source of circulating relaxin throughout pregnancy, the placenta becomes the primary source of progesterone during the last two thirds of human pregnancy (2). In rats, mice, and pigs, a portion of the relaxin accumulates in storage granules within luteal cells until 23 d before delivery (2). The profiles of circulating relaxin, progesterone, and estrogen during pregnancy in rats, pigs, and humans are shown in Fig. 4
. In rats, relaxin is detectable by d 10 of a 23-d pregnancy, and levels rise rapidly to about 100 ng/ml by d 20 (82). Then, during the 23 d before birth, the luteal cells degranulate, and serum relaxin levels surge to maximal levels that frequently exceed 150 ng/ml. This surge in relaxin levels is temporally coincident with the precipitous antepartum decline in progesterone levels that occurs at functional luteolysis and is required for delivery in rats. The general profiles of serum levels of relaxin, progesterone, and estrogen during the 19-d mouse pregnancy are similar to those in the rat (85, 86). The effects of relaxin on reproductive tissues are either dependent upon or augmented by estrogen in rats, mice, and pigs (2). Developing ovarian follicles are the source of rising estrogen levels during late pregnancy in rats and mice (84, 86).
|
The profile of serum relaxin levels during human pregnancy (89, 90) differs markedly from those in the rat, mouse, and pig in three ways (Fig. 3C
). First, the profile of serum relaxin levels follows that of human chorionic gonadotropin: it is highest during the first trimester when the corpus luteum is most active, and it declines throughout the remainder of pregnancy. Second, maximal serum levels of relaxin in humans are only about 1 ng/ml, which is about two orders of magnitude lower than maximal levels in rats and pigs. Finally, because relaxin does not accumulate in human luteal cells as pregnancy progresses (92) and functional luteolysis does not occur, there is no antepartum surge in serum relaxin levels. Whereas the corpus luteum remains the source of circulating relaxin (H2 relaxin) throughout the approximately 40-wk pregnancy (2), the placenta becomes the dominant source of progesterone and estrogen during the second and third trimesters of human pregnancy (2).
Secondary sources of R1, M1, P1, and H2 relaxin that may produce sufficient hormone to act locally in a paracrine manner have been identified in reproductive and nonreproductive tissues (2, 11, 42, 93). Because the well-documented effects of relaxin are attributable to circulating luteal relaxin, potential actions of relaxin in other sources are given limited attention in this review.
With the exception of a few reports of historical importance, all studies of relaxins actions described in this review were conducted by either removing endogenous relaxin in vivo or by administering pure relaxin either in vivo or in vitro. Two pure hormone preparations were administered for essentially all studies. Purified porcine (P1) relaxin (94) was used for in vivo studies in pigs, and rH2 relaxin was used for clinical investigations in humans. Despite the marked differences in their primary structures, both P1 relaxin and rH2 relaxin have high and similar bioactivity in vivo in rodents (2, 18). Moreover, both hormones appear to be equally effective in in vitro systems using either rodent or human cells, tissues, and organs (2, 50, 65, 71, 72, 74, 78, 80, 81).
| II. Effects of Relaxin during Female Reproductive Processes |
|---|
|
|
|---|
A. Follicular growth and ovulation
Whereas relaxin immunoactivity has not been detected in the peripheral blood during the follicular phase of the estrous cycle in any species, relaxin mRNA and/or immunoactivity has been detected in preovulatory follicles in pigs and a variety of other mammalian species including primates (2, 95, 96, 97, 98). Findings in immature pigs led to the hypothesis that relaxin may promote follicular growth through intraovarian autocrine and/or paracrine mechanisms (95, 99). Bagnell et al. (95) reported that the theca interna in prepubertal pigs produces relaxin, and the levels of relaxin in follicular fluid increase with follicular size. Moreover, they found that low doses of P1 relaxin promote growth of both granulosa and theca cells in vitro (100, 101, 102). Consistent with these findings, specific relaxin-binding sites were localized in both the theca and granulosa cells of developing follicles (103).
There is also limited evidence that relaxin may act locally to promote follicular development and ovulation in rats. Proteinase enzymes, including plasminogen activator and collagenase, play an essential role in bringing about the extracellular matrix remodeling required for follicular development and ovulation, and it was demonstrated that P1 relaxin promotes the secretion of these and possibly other proteinases from primary cultures of rat granulosa cells and theca-interstitial cells (104, 105). Additionally, it was reported that rH2 relaxin induces ovulation in the in vitro perfused rat ovary (106), and that passive immunization of circulating R1 relaxin with a monoclonal antibody for R1 relaxin (designated hereafter as MCA1) reduces the number of ovulated oocytes when immature rats are induced to superovulate with gonadotropins (105).
Studies in mice do not support a role for relaxin in either follicular development or ovulation. P1 relaxin did not increase follicular growth or antrum formation in cultures of mouse preantral follicles (107). Moreover, recent studies in M1 relaxin knockout (M1RKO) mice and relaxin receptor LGR7 knockout (LGR7KO) mice provide definitive evidence that M1 relaxin is not required for either follicular development or ovulation in mice. Female mice without either a functional M1 relaxin gene or a functional LGR7 relaxin receptor gene are fertile, and the average litter size does not differ from that of controls (10, 59). A role for M3 relaxin cannot be ruled out because its gene is expressed in ovaries of nonpregnant mice (11). Additionally, one must be mindful that differences exist in the actions of relaxin among species. It remains to be demonstrated that endogenous relaxin plays a role in either follicular growth or ovulation during normal estrous cycles in any species.
B. Implantation
In the 1960s Hisaw and co-workers (108) examined the influence of prolonged administration of progesterone, estrogen, and partially purified P1 relaxin on the histological characteristics of the endometrium in juvenile or ovariectomized rhesus monkeys. The finding that relaxin promoted growth of the endometrium, which included proliferation of the endothelial cells located in the distal portions of the spiral arteries, led these workers to postulate that one of the functions of relaxin is to assist in the preparation of the endometrium for implantation. A recent clinical trial in women with diffuse scleroderma provided additional evidence that relaxin may promote angiogenesis in the human endometrium. During the course of 24 wk of continuous sc infusion, women receiving rH2 relaxin reported heavy, irregular, or prolonged menstrual bleeding more often than women receiving a placebo (109). Because of the interest in the clinical potential of relaxin and the availability of tissue, most of relaxins actions on the endometrium have been determined with in vitro studies using human endometrial cells. Studies that used either P1 relaxin or rH2 relaxin with primary cultures of normal human endometrial cells demonstrated that relaxin binds with high affinity and specificity (110) and increases the expression of hormones, growth factors, and other molecules associated with either decidualization, angiogenesis, or other processes at implantation in humans and other species. Employing cultures of stromal cells, Tseng and co-workers (2, 111) demonstrated relaxin exerts a synergistic effect with progestin on the expression of prolactin, aromatase activity, estrone sulfate sulfatase, and IGF-binding protein-1, which is a major protein synthesized and secreted by decidualized cells. Vascular endothelial growth factor (VEGF), which has been implicated in the new vessel growth and vasodilation that occur in the endometrium at implantation (112), increases with the addition of relaxin to cultures of human endometrium stromal and epithelial cells (109, 113, 114). Glycodelin, a glycoprotein produced and secreted by the secretory endometrium, has been postulated to contribute locally to immunosuppression at implantation (115). Stewart et al. (116) demonstrated that there is a close temporal and quantitative relationship between circulating relaxin and glycodelin profiles during the luteal phase of the menstrual cycle and early pregnancy and that sc administration of rH2 relaxin for 28 d increases the secretion of glycodelin in women demonstrating ovarian cyclicity. It was also reported that P1 relaxin promotes the expression of glycodelin mRNA and protein in primary cultures of human endometrial epithelial cells (117). There is also limited evidence that relaxin increases immunostaining for cyclooxygenase-2 (113) and inhibits the expression of collagenase (114) in primary cultures of human endometrial cells. Relaxin appears to mediate its effects on human endometrial stromal cells, at least in part, through the generation of cAMP (118, 119, 120). There is limited evidence that relaxin acts in concert with estrogen and progesterone to enhance the expression of Hoxa-10, an endometrial transcription factor that is required at implantation in mice (121).
There have been a few in vivo studies of the effects of relaxin on the endometrium in estrogen or estrogen plus progesterone-primed nonpregnant primates and rodents. Porcine relaxin in concert with estrogen was reported to increase endometrial thickness, number of glands, blood vessel content, and VEGF expression in ovariectomized nonpregnant marmosets (122). Similarly, P1 relaxin was reported to promote increased endometrial thickness, loosening of collagen framework, and dilation of blood vessels in rats and mice (123, 124). Finn et al. (125), however, reported that the sc administration of P1 relaxin to estrogen plus progesterone-treated mice does not promote decidualization of the endometrium.
If relaxin plays a role at implantation, it must be available to the uterine endometrium in quantities sufficient to be effective. In humans and macaque monkeys, circulating relaxin is detectable but less than 200 pg/ml at implantation (126, 127). Relaxin levels then rise in close association with chorionic gonadotropin in conceptive cycles (126, 127). A clinical observation makes it seem unlikely that circulating relaxin is important at implantation in humans. Women who experience premature ovarian failure can become pregnant through ovum donation. Implantation occurs in these women despite the fact that they have neither a corpus luteum nor detectable serum relaxin levels (128). The possibility that luteal relaxin contributes to implantation cannot be ruled out for primates. In the marmoset, a New World monkey, serum relaxin levels rise a few days before implantation, and they are higher than in humans and Old World monkeys (129).
It is conceivable that small amounts of relaxin produced locally support implantation through autocrine/paracrine mechanisms. Relaxin mRNA and/or immunoactivity was localized in the endometrium during the luteal phase of the cycle or early pregnancy in humans (114, 130, 131), marmosets (98), pigs (132), rabbits (133), and guinea pigs (134). At present, little is known concerning the amounts, regulation, and function of endometrial relaxin in any species.
As with ovulation, definitive information concerning a putative role for relaxin at implantation comes from findings with both M1RKO and LGR7KO mice. The fact that these animals are fertile, and the average litter size does not differ from that of wild-type controls (10, 59), indicates that M1 relaxin is not essential at implantation in mice. As with ovulation, it remains to be demonstrated that relaxin plays a role at implantation in any species.
C. Effects between implantation and parturition
Studies conducted in pregnant rats and pigs provide convincing evidence that endogenous relaxin produces physiological effects between implantation and parturition. Relaxin affects 1) uterine growth and development, 2) myometrial contractility, 3) central regulation of plasma osmolality, 4) cardiovascular adaptations, and 5) the fetus. With some effects, differences have been demonstrated among species. There is presently no evidence that relaxin is required to maintain pregnancy in any species. The average litter size in rats, mice, and pigs that are devoid of either circulating bioactive relaxin (10, 135, 136, 137) or relaxin receptors (59) during the second half of pregnancy does not differ from that of controls. Also, women who become pregnant with ovum donation maintain their pregnancies (138) despite the fact that relaxin is not detectable in the circulating blood (128). The physiological importance of the effects of relaxin between implantation and parturition remains to be demonstrated.
1. Uterine growth and development.
The administration of P1 relaxin promotes growth of the uterine myometrium and endometrium in nonpregnant rodents and pigs (2, 139, 140, 141, 142). Uterotropic effects of relaxin in nonpregnant rats and/or pigs were reported to be associated with dilation of small arteries and/or veins (123) and increased content of water, protein, collagen, glycogen, DNA, IGFs, IGF-binding proteins, connexins, E-cadherin, VEGF, and tissue inhibitor of matrix metalloproteinases (TIMPs) (2, 142, 143, 144, 145, 146). The uterotropic effects of relaxin are markedly influenced by estrogen. Whereas the administration of P1 relaxin alone over relatively short time periods ranging from 654 h (acute period) promotes growth of the uterus in immature and/or ovariectomized rats and pigs (2, 139, 142, 147, 148), greater growth is obtained when rats are primed with estrogen before relaxin treatment (147). Moreover, when ovariectomized pigs are treated with P1 relaxin for prolonged periods of 10 and 14 d (chronic period), the hormone fails to have a significant effect on uterine growth in the absence of estrogen priming (140, 141). The mechanism(s) whereby relaxin and estrogen act in combination to promote uterine growth remains poorly understood. Estrogen may up-regulate relaxin receptors (2, 148, 149). There is also evidence that the acute effects of relaxin on the rat uterus are mediated through ligand-independent activation of the estrogen receptor (150). It was recently reported that estrogen and relaxin inhibit the expression of estrogen receptor-ß in the rat uterus, and it was postulated that down-regulation of estrogen receptor-ß might be necessary for estrogen or other estrogen receptor activators to exert their full trophic effects on the uterus (151). Progesterone also influences the effects of relaxin on uterine growth in rats and pigs, but the influence of progesterone may differ in the two species. Whereas progesterone inhibited acute relaxin-induced increases in uterine wet weight and collagen content in ovariectomized prepubertal rats (147), progesterone augmented both the acute (139) and chronic (141) uterotropic effects of relaxin in ovariectomized gilts.
The discovery that relaxin has marked effects on growth and development of the uterus in nonpregnant rats and pigs led to the hypothesis that endogenous relaxin plays a role in accommodating the developing fetuses during pregnancy (2). Experimentation did not support the hypothesis in pregnant rats and mice, but did in pigs (Fig. 5
). When circulating R1 relaxin was neutralized with monoclonal antibody MCA1 throughout the second half of rat pregnancy, the uterus was as large at term in relaxin-deficient animals as in controls (152). Consistent with this finding, uterine wet and dry weights increased as dramatically during pregnancy in M1RKO mice as they did in wild-type controls (157). In contrast, when gilts were ovariectomized on d 40 and given hormone replacement therapy with only progesterone, the wet weight of the uterus at term was approximately 30% lower than that in intact controls (135).
|
|
The mechanisms regulating uterine contractility and their hormonal control by relaxin, oxytocin, and other hormones are complex and partially understood (2, 22, 167, 168, 169, 170) (Fig. 7
). It is well known that the signaling cascades that control the concentration of intracellular free calcium (Ca2+) regulate the contractile state of the myometrium. An increase in myometrial cell Ca2+ increases the formation of the Ca2+-calmodulin complex, which then binds to and activates myosin light chain kinase (MLCK). The activated MLCK phosphorylates the 20-kDa regulatory chain of myosin and thereby enhances the interaction of myosin with actin and actin-activated myosin ATPase to bring about contraction. Phosphorylation of MLCK reduces the capacity of MLCK to combine with Ca2+-calmodulin to form the active Ca2+-calmodulin-MLCK complex (2, 167, 168). Relaxin treatment blocks the above cascade of events. After relaxin treatment, rat myometrial cAMP increases, PKA activity increases, affinity of MLCK for the Ca2+-calmodulin complex decreases, MLCK activity decreases, myosin light chain phosphorylation decreases, actomysin ATPase activity decreases, and uterine contractility diminishes (2, 22, 167, 168, 170). Control of phosphorylated light chain/light chain and phosphorylated MLCK/MLCK ratios is complex, and it has not been definitively established how relaxin achieves its effects. The actions of relaxin on myometrial cells are only partially mediated through activation of PKA (2, 22, 167, 170). There is limited evidence that relaxin also acts by reducing intracellular Ca2+ levels by promoting increased Ca2+ efflux and inhibiting mobilization of Ca2+ from intracellular microsomal stores (2, 22, 167, 168). A potential mechanism for promoting hyperpolarization and repolarization that can influence the Ca2+ transient and affect uterine contractility is the opening of K+ channels (22, 168). Sanborn and co-workers (63) demonstrated that relaxin stimulates myometrial Ca2+-activated K+ channel activity and does so through PKA in a human myometrial cell line. There is also limited, but not consistent, evidence that relaxin may stimulate the opening of ATP-dependent K+ channels in isolated rat uterus and myometrium (22, 170). Also inconsistent are reports that relaxin may up-regulate the L-arginine-nitric oxide pathway to increase the second messenger cGMP, thereby inhibiting uterine smooth muscle contractility (2, 80, 170).
|
q)-coupled receptor, thereby activating phospholipase C (PLC). The PLC stimulates phosphoinositol 4,5-bisphosphate turnover and production of inositol triphosphate (IP3) and diacylglycerol. IP3 and diacylglycerol increase intracellular Ca2+ by promoting both Ca2+ release from intracellular endoplasmic reticulum and Ca2+ entry into the cell. Studies that employed P1 relaxin with immortalized pregnant human myometrial cells provide evidence that relaxin inhibits the effects of oxytocin through activation of PKA. After relaxin stimulation, the regulatory subunit of PKA is anchored to the myometrial plasma membrane through association with A kinase anchoring proteins (AKAP), and this obligatory anchoring is required for inactivation of G
q/PLC coupling (171). There is evidence that G
q/PLC inactivation is attributable to phosphorylation of PLCß3 and not G
q (172).
The recently identified relaxin receptor LGR7 (12) is shown to couple directly to a Gs protein to bring about all of relaxins effects through the PKA pathway in Fig. 7
. This is probably an oversimplification. There is evidence that relaxins activation of cAMP in myometrial cells is, at least in part, indirect and through a tyrosine kinase-linked receptor. Kuznetsova et al. (173) reported that the tyrosine kinase inhibitor tryphostin 47 reduced the increase in cAMP content of human myometrium homogenates produced by synthetic H2 relaxin, and Dodge and Sanborn (64) found that the tyrosine kinase inhibitor genistein reversed relaxins effect on oxytocin-stimulated phosphoinositol 4,5-bisphosphate turnover in human myometrial cells. A physiological role of relaxin on human myometrial contractility remains to be demonstrated. Whereas relaxin induces responses in human myometrial cells, there is limited evidence that relaxin does not inhibit contractions of human uterus in vitro (2, 22, 170). The physiological significance of relaxins effects on uterine contractility remains to be demonstrated.
3. Central regulation of plasma osmolality.
Plasma osmolality is maintained within a narrow range by a complex interaction between thirst mechanisms, neurohypophysial release of vasopressin, and the actions of this antidiuretic hormone on the kidney (174). Studies in humans and rats demonstrated that during pregnancy there is an approximately 10 mosmol decline in plasma osmolality that occurs without a change in plasma vasopressin concentrations, and this adaptation of pregnancy is attributable to a reduced osmotic threshold for both thirst and vasopressin secretion (174). There is evidence that relaxin contributes to the decline in plasma osmolality that occurs during pregnancy in rats and mice. Plasma osmolality in nonpregnant rats declined about 10 mosmol after the infusion of either rH2 relaxin or P1 relaxin for 5 or 6 d (19, 175, 176). In pregnant rats, a reduction in plasma osmolality coincides with the elevation in serum relaxin levels during the second half of pregnancy (82, 174). When pregnant rats were made R1 relaxin deficient by either passive immunization with monoclonal antibody MCA1 or ovariectomy, a reduction in plasma osmolality failed to occur (19). Consistent with findings in rats, plasma osmolality during late pregnancy in M1RKO mice is about 10 mosmol higher than in wild-type controls (10). Relaxin may not contribute to the reduction of plasma osmolality that occurs during pregnancy in humans. Plasma osmolality was reported to fall and not differ from that in normal controls in women with singleton pregnancies after ovum donation despite the fact that they have undetectable serum relaxin (128, 177).
a. Drinking.
The effects of relaxin on the osmotic threshold during pregnancy are probably mediated, at least in part, through its central effects on drinking and vasopressin secretion from the posterior pituitary. This hypothesis has been examined only in rats. Water consumption increases markedly during the second half of rat pregnancy (178, 179, 180), and two lines of evidence indicate that relaxin stimulates drinking during this period through actions on the brain. Either intracerebroventricular (icv) or iv administration of relaxin promotes drinking within minutes in nonpregnant rats (180, 181, 182, 183, 184, 185). Second, when passive immunization with monoclonal antibodies for R1 relaxin was used to remove relaxin from either the peripheral circulation (154) or icv fluid (180) throughout the second half of rat pregnancy, water consumption was reduced. Moreover, the reduction in drinking was most profound when relaxin was neutralized in the icv fluid (180).
b. Vasopressin secretion.
Whereas plasma osmolality declines during pregnancy in rats, blood levels of vasopressin do not change significantly (174). Perhaps relaxin acts centrally to contribute to the phenomenon. Either icv or iv administration of P1 relaxin causes secretion of vasopressin in anesthetized nonpregnant rats (15, 186, 187, 188, 189, 190, 191). One can speculate that by acting centrally, endogenous relaxin contributes to the reduced osmotic threshold for vasopressin secretion that occurs during rat pregnancy. By maintaining serum concentrations of vasopressin that are similar to those of nonpregnant rats, water retention is promoted in the renal collecting ducts in pregnant rats despite plasma osmolality that would markedly reduce vasopressin secretion in nonpregnant rats. At this time, however, the central effects of endogenous relaxin on vasopressin secretion in pregnant rats have not been demonstrated.
c. Central control mechanisms.
Relaxin appears to initiate its effects on drinking and vasopressin secretion in rats through actions on the subfornical organ (SFO) and organum vasculosum of the lamina terminalis (OVLT), two circumventricular organs located on the anterior wall of the third cerebral ventricle that lack a blood-brain barrier and are accessible to circulating relaxin (Fig. 8
). High-affinity binding sites for rH2 relaxin were reported in both the SFO and OVLT (192). Additionally, after iv administration of either P1 relaxin or rH2 relaxin, the expression of c-fos increased in groups of neurons located in the more peripheral and dorsal parts of the SFO and in the dorsal cap region of the OVLT (184, 185, 193, 194, 195).
|
4. Cardiovascular adaptations.
During human pregnancy, cardiovascular adaptations that are observed by 58 wk include not only increased plasma volume, cardiac output, and heart rate but also decreased blood pressure and vascular resistance (199, 200). Similar cardiovascular changes occur relatively later in rat pregnancy (178, 201, 202). Recent studies provide evidence that relaxin may contribute to cardiovascular adaptations during pregnancy through effects on the kidney, vasculature, and heart.
a. Glomerular filtration and effective renal plasma flow.
Conrad and co-workers (19) and Conrad and Lindheimer (203) postulated that the circulations of nonreproductive organs such as the kidney serve as arteriovascular shunts that bring about a fall in ventricular afterload during pregnancy. The decrease in ventricular afterload initiates the increase in cardiac output and expansion of plasma volume that occurs during rat and human pregnancy. In pregnant rats, glomerular filtration rate and effective renal plasma flow increase whereas effective renal vascular resistance declines (19), and these renal adaptations are maximal during the second half of pregnancy (204) when serum relaxin levels are elevated. The above renal adaptations of pregnancy failed to occur when circulating R1 relaxin was neutralized with monoclonal antibody MCA1 (19). Relaxin may play a role in these adaptations of pregnancy through direct effects on the vasculature. Conrad and co-workers (18, 19, 205, 206) demonstrated that relaxin is a potent renal vasodilator in rats, promoting both a reduction in myogenic reactivity of small arteries and attenuation of the vasoconstrictive response of the vasculature to angiotensin II. There is evidence that relaxin up-regulates endothelin type B receptors on the vascular endothelium and also the release of nitric oxide, a potent vasodilator (18, 72, 176, 205). Recently, Conrad and co-workers (206) obtained novel evidence that relaxins effects on the renal vasculature are dependent upon increased activity of vascular matrix metalloproteinase (MMP) 2. Findings indicate that this vascular gelatinase acts upstream of, and in series with, the endothelial endothelin type B receptor-NO signaling pathway. It appears to do so by processing big endothelin 1 to endothelin-1132.
b. Vasodilation and blood pressure.
In rats, there is a small reduction in blood pressure during midpregnancy and a larger decline during the 2 or 3 d that precede delivery (204, 207, 208). It is conceivable that endogenous circulating relaxin contributes to this decline in blood pressure. Relaxin has been reported to dilate not only microvessels such as arterioles that are surrounded by a smooth muscle coat but also capillaries and postcapillary venules in numerous sites throughout the body. In rodents, relaxin promotes dilation of microvessels in not only reproductive organs (123, 124, 209, 210) but also in nonreproductive sites, in addition to the kidney described above, including the heart (79, 211, 212), liver (213), and mesocecum (214). Relaxin has also been reported to blunt responses to vasoconstrictors in perfused mesenteric artery of spontaneously hypertensive rats (207, 215), primary cultures of bovine aorta smooth muscle cells (73), rat coronary endothelial cells (74), and segments of rat uterine artery (216). In the primary cell cultures, pretreatment with P1 relaxin markedly reduced the intracellular rise in Ca2+ induced by the vasoactive agonists
-thrombin and/or angiotensin II (73, 74). As in the kidney, relaxin appears to mediate its effects on vasodilation in mesenteric arteries and other sites, at least in part, by acting on endothelial cells to up-regulate endothelin type B receptors and also to stimulate the activity of inducible nitric oxide synthase II, thereby increasing nitric oxide production (18, 72, 74, 79, 205, 211, 216, 217). The connection between these two pathways in relaxin signal transduction, if any, remains to be elucidated.
It is possible that relaxin does not promote vasodilation in all blood vessels. Whereas in vitro treatment with rH2 relaxin was reported to cause a rapid relaxation of human preconstricted gluteal arteries in an endothelium-dependent manner, it did not influence small pulmonary resistance arteries, uterine myometrial arteries, or placental stem villus arteries (217, 218). Moreover, the hormone does not appear to play a role in the antepartum decline in blood pressure in rats. Acute iv administration of relaxin during late pregnancy did not influence blood pressure in conscious rats (219) or anesthetized rats (191). Moreover, removal of endogenous relaxin by either bilateral ovariectomy (208) or by passive immunization of endogenous relaxin (19) failed to influence mean arterial pressure in pregnant rats. Consistent with this finding, short-term infusion of rH2 relaxin had no effect on blood pressure in pregnant rhesus monkeys (220).
c. Heart contractile activity.
The heart is a target organ for relaxin. Sites that bind radiolabeled rH2 relaxin with high affinity and specificity were reported in rat atria (110, 149, 221, 222), and mRNA for the relaxin receptor LGR7 was also identified in the rat (57, 58), mouse (59), and human (12) heart. Numerous in vitro studies have shown that relaxin has potent, direct and concentration-dependent chronotropic and inotropic effects on the isolated rat heart. Relaxin increased not only the rate of spontaneous contractions in perfused intact hearts (211, 223, 224, 225) and isolated right atria (20, 21, 48, 226, 227) but also the force of electrically stimulated contractions in isolated left atria (20, 21, 48, 226, 227). Relaxin may act on both atrial and ventricular pacemakers because relaxin increased heart rate in heart preparations in which the atria had been removed (223). The signal transduction pathways whereby relaxin enhances contractility of the heart have received limited experimental attention. Relaxins chronotropic effects on isolated perfused rat hearts were accompanied by the secretion of atrial natriuretic peptide, and both effects of relaxin appeared to involve cellular signal transduction pathways involving PKA, protein kinase C, and calcium/calmodulin-dependent protein kinases (225). When individual cells were examined with whole-cell patch clamp, relaxin was found to inhibit outward potassium currents, increase action potential duration, and enhance calcium entrance into rat atrial myocytes (228, 229). In similar experiments, relaxin increased the rate of action potentials and L-type calcium current in rabbit sinoatrial node cells (230). These effects involved the activation of PKA (228, 229, 230).
Does endogenous relaxin influence heart contractility in intact pregnant rats? Acute iv or icv administration of P1 relaxin increased heart rate in urethane-anesthetized nonpregnant rats (186, 187, 231), and infusion of rH2 increased heart rate in unanesthetized nonpregnant rats (21). However, acute arterial administration of rH2 relaxin on d 19 did not influence heart rate in conscious pregnant rats (219). The effect, if any, of endogenous circulating relaxin on heart rate in pregnant rats has not been reported.
Interestingly, the central and cardiovascular effects of relaxin appear to be as profound in male as in female rodents. In males, relaxin was reported to decrease plasma osmolality (176) and increase drinking (181, 183, 185, 194), vasopressin secretion (232), glomerular filtration rate, effective renal plasma flow (176), vasodilation (212, 213, 214), and heart rate (20, 21, 223). Thus, unlike the female rodent reproductive tract, relaxins effects on the brain, kidney, general vasculature, and heart do not require elevated circulating estrogen.
5. Fetus.
There is evidence that circulating maternal relaxin influences fetal development during rat pregnancy. When rats were deprived of circulating relaxin throughout the second half of pregnancy by either immunoneutralization of R1 relaxin with monoclonal antibody MCA1 or ovariectomy, fetal weights were significantly greater than in controls (137, 152, 154, 155, 156). It is not known whether this enlargement of fetuses is attributable to the effects of relaxin deprivation on the mother or on the fetus. It does appear likely that only small amounts of maternal relaxin pass to the fetal serum. Whereas serum levels of relaxin in rat fetuses are not known, those in hamster and human fetuses were reported to be low (233) and nondetectable (234, 235), respectively. Consistent with this view are reports of transplacental passage of small amounts of rH2 relaxin in pregnant rhesus monkeys (236, 237). Again, one must be mindful of possible differences among species. An influence of circulating maternal relaxin on fetal weights was not observed with either M1RKO mice (238) or ovariectomized pigs (135). There is also limited evidence that relaxin contributes to testicular descent during late rat pregnancy (239), and that will be described in more detail in Section III.
D. Parturition
1. Time of onset of delivery.
The preponderance of data indicates that circulating endogenous relaxin does not influence the duration of pregnancy. The time of onset of delivery in rats in which circulating R1 relaxin was immunoneutralized with monoclonal antibody MCA1 (240, 241) and in mice that lacked either a functional relaxin gene (10) or relaxin receptor LGR7 gene (59) did not differ from that of controls. There is one report that is not consistent with these findings. When the soluble ligand-binding portion of the human relaxin receptor LGR7 was administered sc to antagonize endogenous circulating relaxin the last 4 d of mouse pregnancy, delivery was delayed by 27 h (12).
Relaxin 1 mRNA and/or immunoreactivity have been reported to be produced within the rat and mouse brain (11, 93, 221, 242), and there is limited evidence that there may be a central R1 relaxin system involving the SFO that influences the time of onset of birth in rats (196, 243). The onset of both luteolysis and delivery was advanced approximately 24 h after central immunoneutralization of relaxin by daily injection of monoclonal antibody for R1 relaxin into the right lateral ventricle throughout the second half of pregnancy (180). It was postulated that central R1 relaxin may influence the time of delivery by acting on the SFO to inhibit oxytocin secretion (180, 196). There are observations that are not consistent with this hypothesis. Intravenous administration of P1 relaxin increased the secretion of oxytocin in unanesthetized nonpregnant and pregnant rats (189, 244). Moreover, neither the time of onset nor the duration of active labor in oxytocin knockout mice differed from those of wild-type controls (245).
2. Duration of delivery and incidence of live young.
Two experimental approaches demonstrated that circulating relaxin has vital effects at parturition in rats and pigs. When primiparous rats and pigs were bilaterally ovariectomized during the second half of pregnancy and given hormone replacement with physiological amounts of progesterone plus estrogen (rats) or progesterone only (pigs), the duration of delivery was several times longer, and the incidence of live births was far lower than in intact controls (136, 244, 246, 247). However, when hormone replacement included physiological levels of P1 relaxin, birth parameters were similar to those of controls (136, 246, 247). Comparable findings were obtained in both rats (240, 241, 248) and pigs (249) when circulating relaxin was immunoneutralized (Fig. 9
).
|
At present, there is no solid evidence that endogenous relaxin facilitates delivery in humans. Circulating levels of relaxin are so low (Fig. 4C
) that they may not be sufficient to do so. The extremely high cesarean section rate of more than 50% in women who become pregnant after oocyte donation and have no functional corpus luteum (128, 250) may be largely attributable to proactive clinical management rather than to an absence of systemic relaxin. Also complicating our understanding of the role of relaxin at term in humans is the possibility that the small amounts of H1 relaxin and H2 relaxin that are produced by the decidua and placenta act through local autocrine/paracrine signaling to increase the expression of MMPs in fetal membranes and thereby bring about their rupture and the induction of delivery (2, 23, 251).
Available evidence supports the view that the primary means whereby relaxin facilitates birth in rats, mice, and pigs is by promoting dramatic growth and remodeling of the cervix. The effects of relaxin on two other portions of the birth canalthe vagina and the interpubic ligamentmay also play roles at the time of delivery in some species.
a. Cervix.
During pregnancy relaxin plays a major role in bringing about the dramatic growth of the cervix that occurs in rats (136, 137, 153, 154, 155, 156, 241, 248, 252), mice (157), and pigs (135, 253) (Fig. 5
). Both the wet weight of the cervix and the circumference of the cervical lumen at term in control rats and pigs are about 2-fold greater than those in relaxin-deficient animals (137, 253) (Fig. 10
, A and B). Numerous and relatively large cervical lumen involutions in relaxin-replete control rats enable far greater expansion of cervical lumens than is possible in relaxin-deficient monoclonal antibody MCA1-treated rats (210). Recently, progress has been made toward an understanding of the mechanisms whereby endogenous relaxin contributes to growth of the rat cervix. Relaxin promotes the accumulation of both epithelial cells and stromal cells (136, 137) not only by stimulating cell proliferation (254) but also by inhibiting apoptosis (255). Within the stroma, relaxins effects on cell proliferation and apoptosis are primarily on fibroblasts and not on smooth muscle cells (254, 255).
|