OPINION

Human Reproduction Vol.19, No.3 pp. 486±489, 2004 Advance Access publication 29 January, 2004

DOI: 10.1093/humrep/deh124

OPINION New insights into the pathophysiology of ovarian hyperstimulation syndrome. What makes the difference between spontaneous and iatrogenic syndrome? A.Delbaere1,6, G.Smits2,3, O.Olatunbosun4, R.Pierson4,5, G.Vassart2,3 and S.Costagliola3 1

Clinique de Fertilite and 2Service de GeÂneÂtique meÂdicale, HoÃpital Erasme, Brussels, 3Institut de Recherche Interdisciplinaire en Biologie Humaine et MoleÂculaire (IRIBHM), Faculte de MeÂdecine, Universite Libre de Bruxelles, Brussels, Belgium, 4Department of Obstetrics, Gynecology and Reproductive Sciences and 5Reproductive Biology Research Unit, College of Medicine, University of Saskatchewan, Saskatoon, Canada 6 To whom correspondence should be addressed at Clinique de FertiliteÂ, HoÃpital Erasme, 808 route de Lennik, 1070 Brussels, Belgium. E-mail: [email protected]

The recent identi®cation of mutations in the FSH receptor gene, which display an increased sensitivity to hCG and are responsible for the development of spontaneous ovarian hyperstimulation syndrome (OHSS), provides for the ®rst time the molecular basis for the physiopathology of spontaneous OHSS. Based on these recent ®ndings, this paper underlines the differences between spontaneous and iatrogenic OHSS and proposes a model to account for the different chronology between the two forms of the syndrome. In the iatrogenic form, the follicular recruitement and enlargement occur during ovarian stimulation with exogenous FSH, while in the spontaneous form, the follicular recruitment occurs later through the stimulation of the FSH receptor by pregnancy-derived hCG. In both forms, massive luteinization of enlarged stimulated ovaries ensues, inducing the release of vasoactive mediators, leading to the development of the symptoms of OHSS. Key words: FSH receptor/OHSS/physiopathology

Ovarian hyperstimulation syndrome (OHSS) is a potentially life-threatening complication of pharmacological ovarian stimulation. Severe forms complicate ~1% of IVF cycles and are characterized by a massive ovarian enlargement together with a ¯uid shift into extravascular compartments responsible for the development of ascites, sometimes pleural and/or pericardial effusion, hypovolaemia, oliguria, and hydroelectrolytic disorders. In the most marked cases, thromboembolic phenomena may occur as a result of haemoconcentration and coagulation disturbances (Hollemaert et al., 1996). It usually involves patients with an explosive response to the ovarian stimulation and is more frequent in patients suffering from polycystic ovary syndrome (Navot et al., 1992). The pathophysiology of the syndrome has not been completely elucidated yet. Exclusively post-ovulatory, the vascular ¯uid leakage is thought to result from an increased capillary permeability of mesothelial surfaces under the action of one or several vasoactive ovarian factor(s) (Elchalal and Schenker, 1997). hCG is thought to play a crucial role in the development of the syndrome: severe forms are indeed restricted to cycles with exogenous hCG (to induce ovulation or as luteal phase support) or with endogenous pregnancy-derived hCG (Delbaere et al., 1997b). 486

Spontaneous forms of OHSS are very rare and always reported during pregnancy. Several cases have been observed during multiple pregnancies (Check et al., 2000) or hydatidiform moles known to be associated with abnormally high values of hCG (Ludwig et al., 1998). Other cases were associated with hypothyroidism and it was suggested that the high levels of thyroid-stimulating hormone (TSH) could stimulate the ovaries (Nappi et al., 1998). A series of cases were recurrent with the development of the syndrome reported in two to six consecutive pregnancies (Zalel et al., 1995; Olatunbosun et al., 1996; Di Carlo et al., 1997; Edi-Osagie and Hopkins, 1997). Spontaneous forms of OHSS were generally reported to develop between 8 and 14 weeks amenorrhoea, differing from iatrogenic OHSS usually starting between 3 and 5 weeks amenorrhoea. We recently identi®ed a mutation in the FSH receptor gene in a patient presenting spontaneous OHSS during each of her four pregnancies (Smits et al., 2003a). The mutation consisted of a substitution of an adenine for a guanine at the ®rst base of codon 567 in exon 10 of the follitropin receptor gene, resulting in the replacement of an aspartic acid with an asparagine. When tested in vitro, the functional response of the mutant receptor displayed an enhanced basal activity and an increased sensitivity to hCG. A distinct mutation in the FSH receptor

Human Reproduction vol. 19 no. 3 ã European Society of Human Reproduction and Embryology 2004; all rights reserved

Spontaneous and iatrogenic OHSS

Figure 1. Chronology of iatrogenic and spontaneous ovarian hyperstimulation syndrome. In the iatrogenic form, the follicular recruitment and growth occur during the administration of exogenous FSH. In the spontaneous form, the follicular recruitment and enlargement occur later through the promiscuous stimulation, by pregnancy-derived hCG, of a mutated FSH receptor (abnormally sensitive to hCG) or a wild type FSH receptor (in the presence of abnormally high levels of hCG). In both forms, massive luteinization of enlarged stimulated ovaries ensues, inducing the release of vasoactive mediators, leading to the development of the symptoms.

gene was concomitantly reported in a patient who developed OHSS during all of her four pregnancies that went beyond 6 weeks of gestation (Vasseur et al., 2003). The mutation was also found in the DNA of two patient's sisters who similarly presented spontaneous OHSS during their pregnancies but not in that of a third unaffected sister. The mutation consisted of a substitution of a thymidine for a cytosine in exon 10 of the follitropin receptor gene resulting in the replacement of a threonine by an isoleucine at position 449 of the follitropin receptor protein. In vitro characterization of the mutated receptor also revealed an increased sensitivity to hCG. In these reported cases, the abnormal functionality of both mutant FSH receptors in vitro provides a straightforward explanation for their implication in the OHSS development in vivo. During pregnancy, the expression of FSH receptor decreases drastically in the corpus luteum, but remains constant in granulosa cells of developing follicles (Simoni et al., 1997). These receptors are usually not or only very weakly stimulated during pregnancy, as pituitary gonadotrophins fall to very low or undetectable levels in serum. The mutated FSH receptor expressed in the developing follicles may be hyperstimulated by the pregnancy-derived hCG. Accordingly, the follicles may start growing, enlarge and ®nally acquire LH receptors on granulosa cells which may also be stimulated by hCG, inducing follicular luteinization together with the secretion of vasoactive molecules responsible for the development of the syndrome.

The interaction between hCG and the FSH receptor could be an essential prerequisite in the development of spontaneous OHSS and could explain why symptoms in spontaneous cases of OHSS appear later than in iatrogenic OHSS in which follicular recruitment and enlargement occur during the ovarian stimulation with exogenous FSH (Figure 1). It is likely that the stimulation of the mutated FSH receptor occurs at a threshold level of hCG which could vary according to the type of mutation. HCG usually peaks between 8 and 10 weeks of pregnancy and declines thereafter. Consequently, the initiation of follicular growth by pregnancy-derived hCG could start between 6 and 10 weeks amenorrhoea. Assuming that these follicles have the same development rate as during a normal or a stimulated cycle (~2 weeks before starting luteinization), the development of the OHSS symptoms is expected to occur in parallel to the massive follicular luteinization, thus between 8 and 12 weeks amenorrhoea (Figure 1). This corroborates most clinical observations as the symptomatology of spontaneous cases of OHSS usually develops as of 8 weeks amenorrhoea culminating at the end of the ®rst trimester of pregnancy. It is noteworthy that experimental follicular maturation and ovulation during pregnancy had been previously observed in primates (Di Zerega and Hodgen, 1979) and in women (White and Bradbury, 1965). Interestingly, the experimental administration of hMG during the third trimester of human pregnancy, 487

A.Delbaere et al.

prior to planned Caesarean section, induced the formation of theca lutein cysts if started 15 days prior to delivery (White and Bradbury, 1965). Taken together, these ®ndings demonstrate that during pregnancy, a cohort of follicles exist which are competent to achieve pre-ovulatory status within 10±14 days and to undergo subsequent luteinization under the action of hCG. It is possible that premature luteinization occurs in some of these secondarily recruited follicles. However, the hormonal environment during pregnancy is very different from that during the follicular phase of a menstrual cycle as high levels of steroids from the placenta (particularly estradiol) may act in synergy with the FSH receptors to promote granulosa cell survival and proliferation. Glycoprotein hormones (hCG, LH, FSH and TSH) are heterodimers consisting of a common a-subunit and a receptorspeci®c b-subunit. The glycoprotein hormone receptors are G protein-coupled receptors characterized by a large extracellular domain responsible for the speci®c recognition and binding of the hormones. HCG and LH, displaying very high sequence similarity between their b-subunits, bind to the same LH/CG receptor, whereas TSH and FSH bind to TSH receptor and FSH receptor respectively. The above-described FSH receptor mutations both induce a reduction of ligand speci®city allowing the activation of the mutated receptor by hCG. Similarly, a mutation in the TSH receptor increasing the af®nity for hCG has also been shown to be responsible for familial gestational hyperthyroidism (Rodien et al., 1998). Spontaneous cases of OHSS are probably not all related to a mutation in the FSH receptor gene. Other clinical conditions could induce the development of the syndrome: a similar mechanism of follicular recruitment, enlargement and luteinization might happen in spontaneous cases of OHSS associated with high levels of hCG. It has been shown recently that hCG was able to stimulate TSH and FSH receptors in vitro in conditions that mimic high ligand concentrations (Schubert et al., 2003; Smits et al., 2003b; Vischer et al., 2003). In vivo, the activation of the TSH receptor by hCG has been demonstrated in cases of gestational hyperthyroidism (Glinoer et al., 1997). Very high concentrations of hCG as found in molar or multiple pregnancies could therefore also stimulate FSH receptors expressed in developing follicles. Altogether, these observations support the development of OHSS, whether iatrogenic or spontaneous, as a consequence of multiple corpora lutea formation associated with active angiogenesis and increased capillary permeability. Luteinization of enlarged superstimulated ovaries may induce the massive release of vasoactive mediators such as vascular endothelial growth factor (McClure et al., 1994), angiotensin II (Delbaere et al., 1994, 1997a) and various interleukins (Abramov et al., 1996), exacerbating local in¯ammatory-like reactions accompanying angiogenesis during corpora lutea formation. As the release of vasoactive molecules reaches a threshold, physiological control mechanisms can be overstretched, leading to the development of the symptoms of OHSS. It is noteworthy that the kinetics of the symptoms are closely related to the lifespan of the corpus luteum. During iatrogenic OHSS, in the absence of pregnancy, symptoms will resolve 488

spontaneously with the onset of the menses; in the presence of pregnancy, symptoms usually start to improve after the sixth week of pregnancy, before the hCG peak. Accordingly, it has been demonstrated that the activity of the corpus luteum diminishes from the ®fth week of pregnancy despite increasing hCG levels (Tulchinsky and Hobel, 1973). During spontaneous OHSS, the `initial' corpus luteum related to the pregnancy is not responsible for the development of the OHSS symptomatology. The formation of `secondary' multiple corpora lutea, or at least of a critical mass of luteinized granulosa cells, could induce a massive release of vasoactive mediators leading to the development of the syndrome. The role of estradiol (E2) levels has been recently debated in the prediction of OHSS (Orvieto, 2003; Aboulghar, 2003). While a threshold level of E2 has been used in several studies to withhold hCG administration during ovarian stimulation for IVF cycles (Delvigne and Rozenberg, 2002; Aboulghar and Mansour, 2003), its reliability has been questioned especially since OHSS also occurs in patients who conceive spontaneously and whose pre-ovulatory serum E2 levels are far less than those encountered after pharmacological ovarian stimulation (Orvieto, 2003). These two pieces of apparently con¯icting clinical evidence can be reconciled considering that in iatrogenic OHSS, high pre-ovulatory E2 levels are the expression of multiple follicular recruitment by exogenous gonadotrophin administration, while in spontaneous OHSS the follicular recruitment occurs later under the action of pregnancy-derived hCG, leading in all cases to massive luteinization of enlarged stimulated ovaries, the ®rst step in the cascade of events leading to the development of OHSS. In conclusion, these recent ®ndings provide for the ®rst time the molecular basis for the physiopathology of spontaneous OHSS and open new perspectives to understand the development of iatrogenic OHSS. While a mutation in the FSH receptor gene should be sought in the presence of recurrent or familial spontaneous OHSS, it would be interesting to ascertain whether polymorphisms of glycoprotein hormone or receptor genes could constitute risk factors in the development of iatrogenic OHSS. References Aboulghar M (2003) Prediction of ovarian hyperstimulation syndrome (OHSS). Estradiol level has an important role in the prediction of OHSS. Hum Reprod 18,1140±1141. Aboulghar MA and Mansour RT (2003) Ovarian hyperstimulation syndrome: classi®cations and critical analysis of preventive measures. Hum Reprod Update 9,275±289. Abramov Y, Schenker JG, Lewin A, Friedler S, Nisman B and Barak V (1996) Plasma in¯ammatory cytokines correlate to the ovarian hyperstimulation syndrome. Hum Reprod 11,1381±1386. Check JH, Choe JK and Nazari A (2000) Hyperreactio luteinalis despite the absence of a corpus luteum and suppressed follicle stimulating concentrations in a triplet pregnancy. Hum Reprod 15,1043±1045. Delbaere A, Bergmann PJM, Gervy-Decoster C, Staroukine M and Englert Y (1994) Angiotensin II immunoreactivity is elevated in ascites during severe ovarian hyperstimulation syndrome: implications for pathophysiology and clinical management. Fertil Steril 62,731±737. Delbaere A, Bergmann PJM, Gervy-Decoster C, Camus M, de Maertelaer V and Englert Y (1997a) Prorenin and active renin concentrations in plasma and ascites during severe ovarian hyperstimulation syndrome. Hum Reprod 12,236±240.

Spontaneous and iatrogenic OHSS Delbaere A, Bergmann PJM, Gervy-Decoster C, Deschodt-Lanckman M, de Maertelaer V, Staroukine M, Camus M and Englert Y (1997b) Increased angiotensin II in ascites during severe ovarian hyperstimulation syndrome: role of early pregnancy and ovarian gonadotropin stimulation. Fertil Steril 67,1038±1045. Delvigne A and Rozenberg S (2002) Epidemiology and prevention of ovarian hyperstimulation syndrome (OHSS): a review. Hum Reprod Update 8,559± 577. DiCarlo C, Bruno PA, Cirillo D, Morgera R, Pellicano M and Nappi C (1997) Increased concentrations of renin, aldosterone and Ca125 in a case of spontaneous, recurrent, familial, severe ovarian hyperstimulation syndrome. Hum Reprod 12,2115±2117. DiZerega G and Hodgen G (1979) Pregnancy-associated ovarian refractoriness to gonadotropin: a myth. Am J Obstet Gynecol 134,819±822. Elchalal U and Schenker JG (1997) The pathophysiology of ovarian hyperstimulation syndromeÐviews and ideas. Hum Reprod 12,1129±1137. Edi-Osagie ECO and Hopkins RE (1997) Recurrent idiopathic ovarian hyperstimulation syndrome in pregnancy. Br J Obstet Gynaecol 104,952± 954. Glinoer D (1997) The regulation of thyroid function in pregnancy: pathways of endocrine adaptation from physiology to pathology. Endocr Rev 18,404± 433. Hollemaert S, Wautrecht J-C, Capel P, Abramowicz MJ, Englert Y and Delbaere A (1996) Thrombosis associated with ovarian hyperstimulation syndrome in a carrier of the factor V Leiden mutation. Thromb Haemost 76,275±276. Ludwig M, Gembruch U, Bauer O and Diedrich K (1998) Ovarian hyperstimulation syndrome (OHSS) in a spontaneous pregnancy with fetal and placental triploidy: information about the general pathophysiology of OHSS. Hum Reprod 13,2082±2087. McClure N, Healy DL, Rogers PAW, Sullivan J, Beaton L, Haning Jr, RV, Connolly DT and Robertson DM (1994) Vascular endothelial growth factor as capillary permeability agent in ovarian hyperstimulation syndrome. Lancet 344, 235±236. Nappi RG, Di Nero E, D'Aries AP and Nappi L (1998) Natural pregnancy in hypothyroid woman complicated by spontaneous ovarian hyperstimulation syndrome. Am j Obstet Gynecol 178,610±611. Navot D, Bergh PA and Laufer N (1992) Ovarian hyperstimulation syndrome in novel reproductive technologies: prevention and treatment. Fertil Steril 58,249±261. Olatunbosun OA, Gilliland B, Brydon LA, Chizen DR and Pierson RA (1996) Spontaneous ovarian hyperstimulation syndrome in four consecutive pregnancies. Clin Exp Obstet Gynecol 23,127±132.

Orvieto R (2003) Prediction of ovarian hyperstimulation syndrome. Challenging the estradiol mythos. Hum Reprod 18,665±667. Rodien P, Bremont C, Sanson ML, Parma J, Van Sande J, Costagliola S, Luton JP, Vassart G and Duprez L (1998) Familial gestational hyperthyroidism caused by a mutant thyrotropin receptor hypersensitive to human chorionic gonadotropin. N Engl J Med 339,1823±1826. Schubert RL, Narayan P and Puett D (2003) Speci®city of cognate ligandreceptor interactions: fusion proteins of human chorionic gonadotropin and the heptahelical receptors for human luteinizing hormone, thyroidstimulating hormone, and follicle-stimulating hormone. Endocrinology 144,129±137. Simoni M, Gromoll J and Nieschlag E (1997) The follicle-stimulating hormone receptor: biochemistry, molecular biology, physiology, and pathophysiology. Endocr Rev 18,739±773. Smits G, Olatunbosun OA, Delbaere A, Pierson RA, Vassart G and Costagliola S (2003a) Spontaneous ovarian hyperstimulation syndrome caused by a mutant follitropin receptor. N Engl J Med 349,760±766. Smits G, Campillo M, Govaerts C, Janssens V, Richter C, Vassart G, Pardo L and Costagliola S (2003b) Glycoprotein hormone receptors: determinants in leucine-rich repeats responsible for ligand speci®city. EMBO J 22,2692± 2703. Tulchinsky D and Hobel CJ (1973) Plasma human chorionic gonadotropin, estrone, estradiol, progesterone, and 17a-hydroxyprogesterone in human pregnancy. Am J Obstet Gynecol 117,884±893. Vasseur C, Rodien P, Beau I, Desroches A, GeÂrard C, de Poncheville L, Chaplot S, Savagner F, Croue A, Mathieu E et al (2003) A chorionic gonadotropin-sensitive mutation in the follicle-stimulationg hormone receptor as a cause of a familial gestational spontaneous ovarian hyperstimulation syndrome. N Engl J Med 349,753±759. Vischer HF, Granneman JC, Noordam MJ, Mosselman S and Bogerd J (2003) Ligand selectivity of gonadotropin receptors. Role of the beta-strands of extracellular leucine-rich repeats 3 and 6 of the human luteinizing hormone receptor. J Biol Chem 278,15505±15513. White CA and Bradbury JT (1965) Ovarian theca lutein cysts. Experimental formation in women prior to repeat cesarean section. Am J Obstet Gynecol 92,973±980. Zalel Y, Orvieto R, Ben-Rafael Z, Homburg R, Fisher O and Insler V (1995) Recurrent spontaneous ovarian hyperstimulation syndrome associated with polycystic ovary syndrome. Gynecol Endocrinol 9,313±315. Submitted on August 28, 2003Accepted on November 11, 2003

489