INTRODUCTION

The peptide hormone relaxin was discovered in 1926 by Frederick Hisaw, who observed that injection of serum from pregnant guinea pigs caused the pubic symphysis of virgin guinea pigs to “relax” (30). Numerous roles have been attributed to relaxin, both during and independent of pregnancy. In fact, relaxin treatment results in biological responses in several very different organ systems, many of which are associated with extracellular matrix remodeling. This review is not about the pleiotropic functions of relaxin as this has been reviewed extensively previously (5, 24, 70); instead, it is an overview of the animal models used to investigate the potential for relaxin as a therapeutic in several human diseases. Specifically, this review discusses if the failure to replicate the many beneficial effects of relaxin in clinical trials could be, in part, due to the animal model used in preclinical studies. It also considers the limitations of animal models that should be addressed in future preclinical studies.

EARLY CLINICAL TRIALS IN THE CERVIX

Cervical ripening is a key process during labor. It involves softening of the cervix (loss of structural integrity and tensile strength) through extensive extracellular matrix remodeling (73). This process is essential for cervical dilation; impaired cervical ripening results in prolonged labor that can lead to further complications. Based on studies in pigs, cows, and rodents (50, 66), early clinical trials used porcine relaxin preparations to assess the possibility that relaxin could be a cervical ripening agent (68). One of these demonstrated in a randomized double-blind placebo-controlled trial that porcine relaxin (2 mg in a viscous gel) instilled in the cervical canal on the evening before the surgical induction of labor improved the mean cervical score compared with the placebo treatment (38). These data provided the proof of concept that relaxin treatment could induce labor through actions on the cervix.

Recombinant human relaxin became available in the 1990s, which was an important advance in furthering the clinical potential of this hormone. The idea that relaxin could supplement or replace current treatment regimens (prostaglandin E₂) to prepare the cervix for delivery was revisited. There were two clinical trials in the 1990s that used 1–4 mg recombinant human relaxin in a methylcellulose gel applied to the posterior vaginal fornix the evening before induction. Women who did not spontaneously go into labor within 15 h of gel application were given the standard induction protocol. Both studies reported that relaxin treatment had no effect as a cervical ripening agent before induction of labor at term (7, 14). A key limitation in these clinical trials was the assumption that if relaxin was applied in a methylcellulose gel in the vagina, the peptide would effectively enter the circulation to reach receptors on the cervix. As serum levels of relaxin (0.4–0.5 ng/ml) did not increase after topical application of relaxin, it was assumed this was not the case (14). This limitation was addressed in the most recent clinical trial that investigated relaxin as a cervical ripening drug (NCT00259103). Pregnant (40+ wk) women planned for induction were treated with recombinant human relaxin intravenously for 24 h. Although the study demonstrated relaxin was well tolerated by women in late pregnancy, it reported that supraphysiological serum levels of relaxin (8–13 ng/ml) did not ripen the cervix in women (81).

An important consideration as to why the effects of relaxin treatment on the cervix in animals have not been replicated in human clinical trials might be explained by the marked differences in the pregnancy physiologies in animals compared with humans. First, there are striking differences in the site of steroid hormone (estrogen and progesterone) production. For example, in rodents, progesterone synthesis by the corpus luteum is essential until term, whereas in humans, progesterone synthesis is largely a function of the placenta from 6 to 8 wk of gestation (41). In effect, the corpus luteum is not required to maintain pregnancy in humans. Serum relaxin concentrations also vary between species. In humans, serum relaxin levels increase before embryonic implantation and peak in the first trimester in pregnant women (53). In contrast, circulating relaxin is first detected midpregnancy in most other species examined, with a peripartum surge in mice, rats, and pigs (66). This surge likely contributes to cervical ripening as well as a loosening of the interpubic ligament. This substantial increase in relaxin concentrations at the end of pregnancy does not occur in humans or other primates (66); this could be due to the diminished role of the corpus luteum in human pregnancy. Therefore, relaxin treatment in the form of a 24-h exogenous peripartum surge in late pregnant women may not have had any effect on the cervix in the clinical trial because unlike many other species, endogenous human relaxin may play a limited role in peripartum cervical remodeling. Alternatively, the 24-h infusion protocol may not have been sufficient to initiate extracellular matrix remodeling. Future clinical trials may have to consider longer durations of relaxin treatment, mindful of the potential off-target side effects of a longer relaxin treatment.

IN VITRO FERTILIZATION: A NEW USE FOR RELAXIN?

Despite the significant advances in fundamental knowledge of conception and early pregnancy physiology, the harsh reality is that approximately half of in vitro fertilization (IVF) treatment cycles are unsuccessful (3). This is a major concern for people wanting to conceive using IVF because treatments are expensive and time consuming. Therefore, improving the outcomes of pregnancies through IVF continues to be a major focus in biomedical research. One strategy is to enhance the maternal uterine environment to improve implantation rates (60). Deficiency in vascular endothelial growth factor-induced angiogenesis is thought to contribute to impaired endometrial receptivity in women with recurrent implantation failure following IVF (13). Interestingly, women with a history of recurrent miscarriage present with low levels of circulating relaxin compared with those without a history of miscarriage (4). Furthermore, early preclinical studies in nonhuman primates concluded that relaxin treatment could play an important role in IVF to stimulate endometrial angiogenesis in early pregnancy and improve the likelihood of implantation (28, 29).

Ovariectomized macaque and rhesus monkeys treated with progesterone and estrogen are often used to investigate aspects of the human menstrual cycle and early pregnancy (28). Additional treatment of these monkeys with relaxin enhanced uterine growth, endometrial lymphocyte number, endometrial vascularization (22, 28), and decidualization (22). Notably, relaxin increased the proliferation of endothelial cells in endometrial blood vessels and the number of arterioles in the endometrium (28). An IVF study in the macaque showed that 21 days of recombinant human relaxin infusion (8 µg·kg⁻¹·day⁻¹) during the peri-implantation period improved the uterine environment for implantation (increased endometrial thickness at days 7 and 28 and increased implantation bleeding compared with day 0), resulting in a slightly higher implantation rate (29). However, the overall rates of successful pregnancy were high in both relaxin- and vehicle-treated groups and reflect the drawbacks of using healthy macaques to study infertility (29). In addition, the normal endogenous levels of relaxin that would have already been circulating in these monkeys could have been a confounding factor. Nevertheless, some of the findings in nonhuman primates have been replicated in nonpregnant women. Continuous subcutaneous recombinant human relaxin treatment (25 and 100 µg·kg⁻¹·day⁻¹) for 24 wk resulted in a marked increase in the proportion of women experiencing menometrorrhagia (irregular prolonged or excessive uterine bleeding), consistent with increased endometrial vascularization (76). Human cell studies indicate that relaxin treatment modulates expression of vascular endothelial growth factor (54, 76) and numerous other angiogenesis-related genes in endometrial stromal cells (42). Moreover, increased in vitro relaxin secretion by human granulosa cell culture in IVF is also predictive of implantation success (69). However, further work is needed to realize full potential of relaxin as a therapeutic in the context of IVF.

Progress in this area of research has been considerably hindered by the lack of an animal model that accurately represents infertility in women, specifically with regard to impaired endometrial receptivity. One scientifically reasonable approach would be to use ovariectomized macaques, similar to those reported in the study of Dallenbach-Hellweg et al. (22), to investigate whether or not relaxin administration could improve implantation rates of embryos fertilized and transferred in a controlled IVF setting as described by Hayes et al. (29). Unfortunately, such a study would be extremely expensive and logistically and perhaps ethically challenging, which is why so much research is conducted in rodents. The dissimilarity in early pregnancy relaxin profiles between humans and rodents suggests that certain relaxin effects on tissue targets during the peri-implantation period in primates may not be observed in rodent models. In fact, providing relaxin to mice early in pregnancy when it is usually not present resulted in resorption of embryos due to its inhibition of myometrial activity, which is essential to normal embryo spacing in that species (68). This reinforces the importance of selecting the right animal model to investigate specific relaxin physiologies and disease conditions.

MORE THAN JUST A HORMONE OF PREGNANCY

During normal pregnancy, remarkable changes occur to the maternal cardiovascular system, including an increase in stroke volume (SV), cardiac output (CO), and glomerular filtration rate with a concomitant decrease in systemic vascular resistance (SVR) and mean arterial pressure (1, 16, 39, 40). Multiple studies in rodents demonstrate that relaxin plays an essential role in modulating these profound changes (19).

Relaxin treatment in nonpregnant animals mimics many of the cardiovascular changes that occur in pregnancy. These effects can be replicated in healthy male and nonpregnant female rodents. Briefly, chronic relaxin treatment increases CO, SV, heart rate, renal blood flow, and glomerular filtration rate while decreasing SVR (20). Acute intravenous treatment (2–4 h) with recombinant human relaxin increased CO, SV, and heart rate and reduced SVR in an angiotensin II-induced rat hypertension model, but a longer treatment duration (24 h) was required to see similar effects in spontaneously hypertensive rats (23), highlighting the importance of selection of animal model. A hemodynamic study conducted in patients with acute heart failure demonstrated a rapid (2 h) relaxin-induced decrease in SVR and blood pressure, a trend toward increased cardiac index, and no effect on heart rate (56). Although the hemodynamic studies in both rodents and humans, along with observations made in other clinical trials, support the conclusion that relaxin can cause rapid vasodilation (74), the differences highlighted in the hemodynamic studies reflect the limitations of using a single animal model to try to understand human pharmacology of relaxin. The vascular effects of relaxin, both with regard to vasodilation and its ability to reduce vascular stiffness, have been researched extensively (34), and it became apparent that relaxin only had effects in certain vascular beds (renal, mesenteric, aorta) and not others (femoral, middle cerebral artery) (31). Furthermore, discrepancies occurred between studies in which different regimens of relaxin treatment or different strains, ages, and genders of rodents were used. For example, 3 and 5 days of subcutaneous relaxin treatment reduced mesenteric artery stiffness in male Wistar rats (33, 37) but not in female Wistar Hannover rats (77). More recent work demonstrated that 2 and 10 days of relaxin treatment in male Wistar rats also had no effect on mesenteric artery stiffness (31, 32). The discrepancies between studies raise two important questions: 1) which experimental model in healthy animals best mimics the human, and 2) is it relevant to treat rodents (and other animals) with recombinant human relaxin? As this is the most readily available form of relaxin, it is the most commonly used in experiments, but it is important to consider species specificity, as well as potential issues with immunogenicity when administering human relaxin to animals (59). Regardless, a major limitation of relaxin treatment studies in healthy animals is that they bear little relevance to human disease.

RELAXIN TREATMENT IN ACUTE HEART FAILURE

The numerous studies that demonstrated cardioprotective effects of relaxin gave rise to the development of relaxin as a potential therapeutic in cardiovascular disease and led to large scale clinical trials for the use of relaxin in acute heart failure (25). The phase II clinical trial and first phase III trial (RELAX-AHF-1) yielded promising results and indicated that continuous intravenous recombinant human relaxin treatment (30 µg·kg⁻¹·day⁻¹) for just 48 h improved one of the primary dyspnoea end points in the short term and reduced all-cause mortality at 180 days in acute heart failure patients with normal-to-elevated blood pressure (72). Relaxin also improved renal function as measured by a number of clinical assessments (48). However, there were questions raised around secondary end points of cardiovascular death and readmission to hospital for heart failure so these findings were not sufficient for immediate Food and Drug Administration approval on a single phase III trial. Therefore, a second phase III study (RELAX-AHF-2) was conducted (NCT01870778).

Unlike the earlier phase III trial, early reports indicate that relaxin did not reduce cardiovascular death through day 180 or reduce worsening heart failure in the second phase III trial (71). Attempts to understand the underlying basis for observations in the acute heart failure clinical trials have been limited due to the complex pathophysiology of the syndrome and perhaps the naiveté of drug (47), as well as to the lack of a robust animal model of this disease. Some might argue that relaxin, a molecule with a short half-life and associated limitations on clinically feasible treatment duration, is better suited to treating disorders of acute onset and recovery, instead of attempting to alter an end point remote from the time relaxin was actually administered. The 48-h intravenous infusion utilized in the acute heart failure trials has been shown to rapidly improve hemodynamics (59), consistent with observations in some preclinical models (23), suggesting that relaxin may be better suited to therapeutic indications in which such rapid modulation is clinically meaningful.

Because there are few robust animal models of acute cardiovascular disorders, the majority of relaxin research (and demonstrated efficacy) has been performed in animal models of chronic cardiovascular diseases. These models should be evaluated with regard to their applicability to human disease, and in concert, investments in developing stable small molecule RXFP1 agonists and/or relaxin analogs with an extended half-life should be considered. The majority of animal studies that have revealed cardioprotective effects of relaxin have used treatment durations that exceed 48 h (25, 35). Treatment durations longer than 48 h have not been explored clinically in heart failure due to the fact that relaxin must be administered parenterally and continuously, which results in a relatively rapid rise and stable elevation in serum relaxin levels. The inability of the peptide to be taken orally (as it is broken down during digestion and not absorbed into the circulation) may also be a key limitation in the therapeutic development of relaxin. Oral medications are theoretically more easily tolerated by patients and physicians and are often far cheaper to administer. The therapeutic potential of relaxin would likely increase significantly if this obstacle could be overcome.

The effects of relaxin treatment in chronic cardiovascular diseases (and associated risk factors) have been studied in a number of animal models. These include models of hypertension (23, 64, 82), obesity (78), aging (77), and diabetes (51). Although relaxin has beneficial effects on the vasculature in many of these models (34), there are discrepancies between studies, which are summarized in Table 1. These discrepancies have likely impeded the development of relaxin as a therapeutic in cardiovascular disease.

FUTURE OF RELAXIN: ISCHEMIA-REPERFUSION INJURY

The previously mentioned failures to translate findings from animal studies to clinical trials could indicate that the potential of relaxin as a therapeutic is limited. However, relaxin has undeniable cardioprotective effects in animal models and is particularly effective in the vasculature. Furthermore, the potential benefits of relaxin treatment in other acute cardiovascular disorders have not been explored extensively. There are considerable data to support the use of relaxin in ischemia-reperfusion injury (17, 83). Blockage of blood flow as a result of atherosclerosis and thromboembolism leads to ischemia, tissue necrosis, and organ failure. Current therapeutic strategies are aimed at removing these blockages to reestablish blood flow to the previously ischemic area and limit tissue damage. However, reintroduction of blood flow results in another unique form of tissue damage known as reperfusion injury (58). The effects of relaxin in ischemia-reperfusion injury have been interrogated using numerous in vivo and in vitro animal models of myocardial ischemia-reperfusion injury (6, 43). Relaxin has cardioprotective properties in myocardial infarction-reperfusion injury in pigs (52, 55), rats (6, 12), and guinea pigs (43). In the setting of myocardial ischemia-reperfusion injury, relaxin exerts cardioprotective effects by increasing coronary flow to the ischemic region while preventing cellular damage and death. The fact that relaxin has shown beneficial effects in ischemia-reperfusion models in multiple species is promising. However, translation to clinical use would be significantly enhanced if preclinical studies were undertaken to test the effect of relaxin on a background of currently used medical therapy and included clinically relevant end points, such as mortality (8).

The protective effects of relaxin in myocardial ischemia-reperfusion injury prompted numerous studies in rats to investigate the potential of relaxin to protect against ischemia-reperfusion injury in other organs such as kidneys (17, 83), lungs (2), intestines (44), and brain (11). Collectively, these studies demonstrate that relaxin protects these organs by reducing apoptosis, inflammation, and oxidative stress. This is likely mediated by a mechanism involving the nitric oxide pathway. Ischemia-reperfusion injury is a severe unavoidable consequence of organ transplants, particularly in deceased-donor kidney transplants (57, 61).

Ischemia-reperfusion injury complicates transplants in two steps. First, there is immediate damage caused from the reperfusion during transplantation. Second, the injury activates the innate and adaptive immune response, which increases the risk of rejection (26). There is currently a high demand for therapeutics that aid in reducing ischemia-reperfusion injury in kidney transplants, and relaxin may be a potential solution for this. A number of animal models for kidney transplantation (in dogs, rodents, and more recently pigs) are available, and while they do require skilled microsurgical skills, they are possible (65). These models exhibit similar autoimmune responses to those of humans and allow for a greater understanding of the mechanisms of relaxin-action in ischemia-reperfusion injury in the context of kidney transplants.

NEW USES FOR RELAXIN IN PREGNANCY

Relaxin has also recently been considered as a therapeutic in pregnancy-related hypertension, largely because of its vasoprotective functions. A particularly severe form of hypertension in pregnant women is preeclampsia, which presents as the new-onset of high blood pressure after 20 wk of pregnancy, coupled with other symptoms such as proteinuria and fetal growth restriction. Preeclampsia can quickly escalate into HELLP (hemolysis, elevated liver transaminases, low platelets) syndrome or eclampsia. These serious complications are life threatening for both mother and baby. There are limited treatment options for preeclampsia, and the only cure is delivery of the placenta (49). Unfortunately, when preeclampsia is severe, this often results in extremely premature babies at risk of developing further complications. Although the placenta is central to disease development, it is maternal systemic and renal vasoconstriction that leads to hypertension and impaired renal function. Relaxin treatment could potentially counteract these negative characteristics by targeting the endothelial dysfunction to help stabilize systemic vascular function (21, 75). This rationale was used as the basis for initiating two exploratory clinical trials in which women with preeclampsia were treated intravenously for 72 h with recombinant human relaxin (15 µg·kg⁻¹·day⁻¹; clinicaltrials.gov NCT00333307 and NCT01566630). The objectives were to assess the safety of relaxin and its effects on blood pressure and other symptoms of preeclampsia. Both trials were closed due to the difficulty in recruiting patients.

The challenge in enrolling patients into the trials reflects the need for further rigorous preclinical testing of relaxin in preeclampsia. The main limitation to achieving this goal is the lack of an appropriate animal model of preeclampsia that encompasses all elements of the disorder (46). Models of preeclampsia exist, but there are limitations to each of them, aside from the large differences in placentation between animals and human. The reduced uterine perfusion pressure (RUPP) model is a promising model if the main purpose of the study is to investigate drug effects on the vasculature and renal function, i.e., downstream of placentation. This model presents with characteristics that are observed in women with preeclampsia (hypertension, systemic and renal vasoconstriction, oxidative stress, and intrauterine growth restriction). The model has been successfully implemented in a number of species (27), but it is most consistent with human preeclampsia in rats and primates (36). The primate model most closely resembles the human disease, but as noted above, primate studies have their own set of limitations; therefore, studies in rats are preferred. However, there is a major consideration in the RUPP model. RUPP is generally applied in midpregnancy after the placenta is established. Unfortunately, this prevents the study of abnormal placentation, which leads to placental ischemia in early pregnancy in preeclamptic women. Nevertheless, being able to treat the symptoms of the disorder in mid- to late-pregnant women is a significant clinical goal. A recent study indicated that treatment with relaxin in the rat RUPP model improves a number of these symptoms (62). Specifically, relaxin treatment reduced blood pressure and uterine artery resistance, increased nitric oxide bioavailability in the kidney, and improved circulating levels of preeclamptic biomarkers (sFlt-1, TNF-α, and PPET-1). However, it is yet to be determined whether or not relaxin improves maternal systemic vasculature function.

BREAKING RESEARCH: POTENTIAL OF RELAXINAS A THERAPEUTIC FOR LIVER DISEASE

Over a decade ago it was discovered that relaxin receptors (RXFP1) are expressed in cirrhotic livers but not healthy liver in rats (9). This brought to light the hypothesis that relaxin may be therapeutically useful in liver injury. In subsequent studies, chronic relaxin treatment significantly reduced fibrosis in mouse models of established hepatic fibrosis (10). Recent preclinical studies in rat models of cirrhosis also demonstrated a therapeutic potential for relaxin (67). These studies targeted hepatorenal syndrome (a common complication of cirrhosis) and focused on the ability of relaxin to enhance renal flow and vasodilation. The rat studies translated well to humans, with comparable findings in a phase II clinical trial (67). Patients with alcohol-related cirrhosis were treated intravenously with recombinant human relaxin (80 µg·kg⁻¹·day⁻¹ for 60 min, followed by 30 µg·kg⁻¹·day⁻¹ for 60 min). Relaxin infusion increased total renal arterial blood flow by 65% from baseline in patients and was not associated with any detrimental effects on systemic blood pressure or hepatic perfusion. Importantly, this exciting new evidence highlights that the therapeutic potential of the vasodilatory actions of relaxin is not limited to cardiovascular disease.

GRANTS

M. Jelinic and S. A. Marshall received an Australian Postgraduate Award. C. H. Leo received the JN Peter’s Research Fellowship and an Early Career Grant (Faculty of Science, The University of Melbourne). The research was funded by an Australian Research Council Linkage Grant, the National Health and Medical Research Council, and Investigator-Initiated Trials from Novartis Pharmaceuticals Australia.

DISCLOSURES

We disclose that some of this research (described in this review) was partially funded by Novartis Pharmaceuticals, who also provided the relaxin as a condition of the Australian Research Council Linkage Grant. L. J. Parry was also a paid consultant for Novartis Pharmaceuticals and is a coinventor on a patent for relaxin use in the cervix, kidney, and brain. D. Stewart and E. Unemori are former employees of Corthera, a Novartis Pharmaceuticals company, and inventors on relaxin use patents.

AUTHOR CONTRIBUTIONS

M.J., L.J.P., and C.H.L. drafted manuscript; M.J., S.A.M., D.S., E.U., L.J.P., and C.H.L. edited and revised manuscript; M.J., S.A.M., D.S., E.U., L.J.P., and C.H.L. approved final version of manuscript.