isoforms are produced from this gene by alternate slicing to form active
disul´¬üde-linked homodimers (Keck et al., 1989; Leung et al., 1989; Tischer
et al., 1989).
The ´¬ürst human VEGF were cDNAs cloned from a phorbol ester-activated
HL60 promyelocytic leukemia cell library (Leung et al., 1989) and histiocytic
lymphoma cell line U937 (Connolly et al., 1989). Both the cDNAs were
screened with oligonucleotides designed on the basis of the amino-acid
sequence of the previously puri´¬üed protein. The VEGF family (Figure III.2)
includes four different dimeric forms (A├ÇD) and placental growth factors,
which all bind differently to the three receptors (VEGF-R 1├Ç3) that are
expressed on endothelial cells (Neufeld et al., 1999).
Tyrosine Kinase Receptors
Two VEGF receptors belonging to the tyrosine kinase receptor family have been
identi´¬üed and cloned: the VEGFR-1 and the VEGFR-2 receptors. The third
VEGF receptor, VEGFR-3 receptor, is expressed in lymph vessels and binds
VEGF-C and VEGF-D. These three receptors form a subfamily characterized by
the presence of seven immunoglobulin-like loops in their extracellur part and
a split tyrosine├Çkinase domain in their intracellular part. The homologous
tyrosine kinase receptors, fms-like tyrosine kinase receptor VEGFR-1 (´¬‚t), and
kinase insert domain-containing receptor VEGFR-2 (KDR), function as high
af´¬ünity VEGF receptors (Millauer et al., 1993). KDR and ´¬‚t are selectively
expressed by vascular endothelial cells (Kendall and Thomas, 1993). VEGFR-1
is also expressed in the trophoblast cells, monocytes and mesangial renal cells.
VEGFR-2 is also expressed in the hematopoietic stem cells, megakaryocytes and
retinal progenital cells (Neufeld et al., 1999). The expression of VEGFR-1 and
VEGF GENE EXPRESSION
Fig. III.2: Growth factors and receptors of the VEGF family
Reproduced with permission from Neufeld et al. (1999). FASEB Journal 13:9├Ç22
VEGFR-2 is affected by hypoxia, but to a lesser degree than for VEGF (Neufeld
et al., 1999).
VEGF165 Specific Receptor Neuropilin-1
Endothelial cells also contain VEGF receptors with a lower mass than VEGFR-1
or VEGFR-2 (Gitay-Goren et al., 1992). It was subsequently discovered that
these smaller receptors of the endothelial cells are isoform-speci´¬üc, and bind to
VEGV165 and not to VEGF121. The binding of VEGF165 to these receptors is
mediated by amino acids residing at the carboxy terminal part of the exon 7
encoded peptide of VEGF165. Two such receptors have been identi´¬üed:
neuropilin-1 and neuropilin-2. Neuropilin-1 also functions as a receptor for the
heparin-binding form of placental growth factor. The neuropilins have a short
intracellular domain and therefore cannot function as independent receptors
(Figure III.2). No responses to VEGF165 were observed when cells were
expressing neuropilin-1 but no other VEGF receptors were stimulated with
VEGF165 (Soker et al., 1998).
VEGF GENE EXPRESSION
The human VEGF gene has been mapped to chromosome 6p12 and is made up
of eight exons. Exons 1├Ç5 and 8 are always present in VEGF mRNA, whereas
54 PATHOPHYSIOLOGY OF OVARIAN HYPERSTIMULATION SYNDROME
the expression of exon 6 and 7 is regulated by alternative splicing. The
phenomenon produces various VEGF isoforms which differ in length but have
a common region.
REGULATION OF VEGF PRODUCTION
VEGF production is upregulated by hypoxia, cytokines and prostaglandins
(Ferrara et al., 1992; Rizk et al., 1997). Cytokines and growth factors that do
not stimulate angiogenesis directly can still modulate angiogenesis by impacting
VEGF expression in certain cell types (Neufeld et al., 1999). In other words,
cytokines may have an indirect angiogenic or anti-angiogenic effect. Growth
factors that potentiate VEGF production include transforming growth factor b,
´¬übroblast growth factor-4, platelet derived growth factor, insulin-like growth
factor-1, interleukin-1b and interleukin-6 (Goad et al., 1996; Li et al., 1995;
Cohen et al., 1996; Neufeld et al., 1999). VEGF production is downregulated by
thrombospondin, hyperoxia and interleukin-10 (Enskog et al., 2001a, b). VEGF
is produced and stored in granules in T-lymphocytes, mast cells, neutrophils
VEGF-A, commonly known as just VEGF, exists in at least ´¬üve isoforms of
different molecular weights. Five human VEGF mRNA species encoding VEGF
isoforms of 121, 145, 165, 189 and 206 amino acids (VEGF121├Ç206) are
produced by alternative splicing of the VEGF mRNA. The major difference
between the different VEGF isoforms is their heparin and heparin-sulfate
binding ability. The most potent and well-characterized isoform is VEGF165.
It is made up of two subunits of 165 amino acids. The three secreted
VEGF splice forms, VEGF121, VEGF145 and VEGF165, induce proliferation of
endothelial cells and in vivo angiogenesis.
Role of VEGF in Physiological and Pathological States
In vivo, VEGF is a powerful mediator of vessel permeability. It is also
strongly implicated in the initiation and development of angiogenesis in
thedeveloping embryo and in adult tissue undergoing profound angiogenesis,
such as cycling endometrium and the leuteinizing follicle (Charnok-Jones
et al., 1993). In addition to its physiological role, VEGF is implicated as a
critical angiogenic factor in the development of tumor vascularization (Kim
et al., 1993) and the excessive neovascularization seen in conditions such as
rheumatoid arthritis (Koch et al., 1994). Its levels are also increased in the
peritoneal ´¬‚uid of women with endometriosis compared with normal
controls (Rizk and Abdalla, 2003; McLaren et al., 1996).
VEGF AS A CAPILLARY PERMEABILITY AGENT IN OHSS
Role of VEGF in Reproductive Function and Ovarian Cyst Formation
VEGF may also play a role in the regulation of cyclic ovarian angiogenesis, and
its ability to increase vascular permeability may be an important factor in the
production of Fallopian tube ef´¬‚uent and ´¬‚uid formation in ovarian cysts.
Gordon et al. (1996) demonstrated that, in normal ovaries, VEGF within
healthy follicles was localized to the theca cell layer with minimal VEGF peptide
detected in the granulosa cell layer. VEGF was not expressed in atretic follicles
or degenerating corpus luteum. However, intense VEGF immunostaining was
observed within the highly vascularized corpora luteum. In normal ovaries
from postmenopausal women, VEGF was detected only in epithelial inclusion
cysts and serous cystadenoma. The authors concluded that, during reproductive
life, VEGF plays an important role in growth and maintenance of ovarian
follicles and corpus luteum by mediating angiogenesis. In addition, VEGF
within the fallopian tube luminal epithelium increased the vascular per-
meability and modulated the tubal luminal secretions. Similarly, VEGF in the
epithelial lining of benign ovarian neoplasms may contribute to ´¬‚uid formation
in ovarian cysts.
WHAT IS THE LINK BETWEEN VEGF AND OHSS?
The role of VEGF in OHSS has been extensively evaluated (Aboulghar et al.,
1996; Rizk et al., 1997; Pellicer et al., 1999). The association will be discussed
in this section but it has been very well summarized by Albert et al. (2002), who
suggested three important reasons af´¬ürming the role of VEGF as a potential
mediator in the development of OHSS. First, VEGF and its isoforms have
vasoactive properties (Senger et al., 1983; Motro et al., 1990); second, VEGF has
been identi´¬üed in follicular ´¬‚uid; and third, mRNA transcripts and proteins
have been detected in granulosa luteal cells (Yan et al., 1993; Neulen et al.,
1995a, b; Gordon et al., 1996). Finally, VEGF is increased in serum and
peritoneal ´¬‚uid of women who develop OHSS compared with control patients
(Abramov et al., 1997; Agrawal et al., 1998, 1999; Artini et al., 1998; Ludwig et
al., 1999; Aboulghar et al., 1999).
VEGF AS A CAPILLARY PERMEABILITY AGENT IN OHSS
McClure et al. (1994) pioneered investigation of the role of VEGF as the
capillary permeability agent in OHSS. Two similar peaks of permeability
activity were observed in OHSS ascites and liver ascites spiked with human
VEGF (rhVEGF). No activity was observed in control liver ascites. Incubation
with rhVEGF antiserum decreased activity in the two OHSS peaks by 79% and
65%, and in the two spiked liver peaks by 49% and 50%. In contrast, control
serum produced 24% and 27%, and 17% and 0% reductions, respectively.
56 PATHOPHYSIOLOGY OF OVARIAN HYPERSTIMULATION SYNDROME
These results have led investigators to conclude that VEGF is the major
capillary permeability agent in ascites ´¬‚uid (McClure et al., 1994).
VEGF mRNA Expression in the Rat and Primate Ovary
Molecular biology studies suggest a strong link between VEGF and hCG, which
is important in the development of VEGF (Rizk and Nawar, 2004). First,
hybridization studies demonstrated VEGF mRNA expression in the rat (Phillips
et al., 1990) and primate ovary (Ravindranath et al., 1992) predominantly
after the LH surge. This surge is also essential for OHSS. Second, luteal
phase treatment with GnRH antagonist to suppress LH secretion decreased
VEGF mRNA expression, implying such expression is dependent on LH
(Ravindranath et al., 1992). Similarly, luteal phase supplementation with
progesterone rather than hCG decreases the likelihood of OHSS (Smitz et al.,
1990; Rizk and Smitz, 1992; Novella-Maestra et al., 2005).
Low-dose LH Decreases VEGF Expression in Superovulated Rats
The administration of a combination of pregnant mare serum gonadotrophins
(PMSG) and hCG in high doses induces OHSS, which is characterized by
increased vascular permeability and overexpression of VEGF in the ovarian
cells. It is established that hCG has a longer half-life than LH and a greater
biologic activity expressed in a higher incidence of OHSS. FSH may also be
related to the ovulatory changes within the follicle, based on the fact that there
are cases of spontaneous LH surges without the administration of hCG or LH.
Gomez et al. (2004) compared the capacity of hCG, LH and FSH to induce
ovulation and simultaneously prevent OHSS in the animal model. Immature
female rats were given PMSG (10 IU) for four days. Ovulation was triggered by
using 10 IU of hCG, 10 IU FSH, 10 IU LH, 60 IU LH or saline. The number of
oocytes ovulated into the tubes, vascular permeability and mRNA VEGF
expression were evaluated and compared. All the hormones utilized in this
investigation were equally effective in triggering ovulation, with similar
signi´¬ücant p values when compared with saline controls. The administration
of 10 IU of LH resulted in signi´¬ücantly lower vascular permeability and VEGF
expression than that observed in the groups treated with 10 IU of hCG, 10 IU of
FSH or 60 IU of LH. The authors concluded that FSH and hCG, as well as
a six-fold increase in LH, demonstrated similar biologic acitivities, including
increased vascular permeability, such as VEGF expression (Gomez et al.,
2004). In fact, the lower doses of LH produced similar ovulation rates but,
at the same time, prevented the undesired permeability changes and perhaps
the risk of OHSS.
VEGF m-RNA Expression in Human Luteinized Granulosa Cells
Yan et al. (1993) were the ´¬ürst to demonstrate the presence of VEGF mRNA in
human luteinized granulosa cells. Neulen et al. (1995a, b), from the same
VEGF AS A CAPILLARY PERMEABILITY AGENT IN OHSS
group, later demonstrated that the expression of VEGF mRNA is enhanced
by hCG in a dose- and time-dependent fashion. VEGF mRNA expression in
granulosa cells was enhanced by increasing amounts of hCG, with maximum
enhancement at 1 IU of hCG/ml of medium. Further dosage increments
revealed no additional augmentation of VEGF expression. VEGF mRNA
expression also reached maximum values at 3 h. Kamat et al. (1995) used
immunohistochemistry to demonstrate the increased activity of VEGF with
Graa´¬üan follicle development, which reaches strong cytoplasmic staining for
VEGF with the formation of the corpus luteum.
Is hCG-induced OHSS Associated with Upregulation of VEGFG?
Wang et al. (2002) investigated whether the effects of hCG on the pathogenesis
of OHSS were mediated through the VEGF produced by luteinized granulosa
cells. They measured estradiol, VEGF, and insulin-like growth factor II (IGF-II)
in serum and follicular ´¬‚uid, and analyzed mRNA expression in luteinized
granulosa cells obtained from 101 women (58 with OHSS and 43 controls) who
underwent IVF/ET. HCG upregulated VEGF expression of granulosa cells in
the OHSS and not in the control group. Follicular VEGF worked through an
autocrine mechanism using its kinase insert domain-containing receptor. The
authors calculated total follicular production of VEGF by multiplying follicular
concentrations by volumes. They veri´¬üed that an increase in total follicular
production of VEGF accounted for elevated serum levels of VEGF, which was
associated with the development of OHSS.
VEGF Dynamic Studies and OHSS
A large series of investigations have been completed between 1995 and 2005
that address the relation between VEGF and OHSS (Delvigne, 2004; Rizk and
Nawar, 2004; Rizk and Aboulghar, 2005). While many studies have reported the
correlation between OHSS and serum/plasma, peritoneal ´¬‚uid/follicular ´¬‚uid
VEGF levels, others have reported contradictory results with no difference
between the OHSS and the control groups (Geva et al., 1999; DÔÇ™Ambrogio et al.,
1999; Enskog et al., 2001a). Delvigne (2004) explained several potential
mechanisms for the contradictory results in the large number of VEGF
(1) VEGF can be measured in plasma or in serum but the clotting process
increases VEGF 8- to 10-fold in serum.
(2) Degranulation or hemoconcentration in OHSS may cause misinterpreta-
tion of the real level of free, active VEGF.
(3) VEGF could be trapped in the ascites ´¬‚uid and the large corpora luteal cysts
of the ovary.
(4) The biologically active isoform of VEGF could vary from one patient to
58 PATHOPHYSIOLOGY OF OVARIAN HYPERSTIMULATION SYNDROME
(5) Immunoassays may not be able to differentiate between the four or more
isoforms of VEGF.
(6) Soluble VEGF receptors may in´¬‚uence the biologic activity of VEGF.
Krasnow et al. (1996) measured VEGF in serum, peritoneal ´¬‚uid and
follicular ´¬‚uid of eight patients considered at risk of OHSS. Serum VEGF
was signi´¬ücantly higher in the group who developed severe OHSS compared
with those who did not. The detection of high serum VEGF levels in the
circulation of patients with OHSS suggests that this factor may play a role in the
pathogenesis of OHSS. The large amount of VEGF in follicular ´¬‚uid relative
to serum or peritoneal ´¬‚uid suggests that the ovary is a signi´¬ücant source of
VEGF. In an unstimulated menstrual cycle, the development of a single corpus
luteum does not result in OHSS. In patients with severe OHSS in whom VEGF
was signi´¬ücantly higher in the serum, a mean of 21 follicles were present before
hCG administration. It is possible that the hCG that rescues the corpus luteum
results in an increase in ovarian VEGF secretion, which in turn causes an
exacerbation of OHSS, con´¬ürming the work of Neulen et al. (1995a, b). The
effect of follicular aspiration on the incidence of OHSS has been debated in
clinical studies (Rizk, 2001, 2002).
Abramov et al. (1996) followed the kinetics of VEGF in the plasma of seven
patients with severe OHSS from the time of admission to the hospital until
clinical resolution. High levels of VEGF were detected in the plasma of all
patients admitted for severe OHSS compared with controls, who received
similar ovulation-induction regimens but did not develop OHSS after IVF
and embryo transfer. Levels dropped signi´¬ücantly, accompanied by clinical
improvement, reaching minimum values after complete resolution. A
statistically signi´¬ücant correlation was found between plasma VEGF levels
and certain biological characteristics of OHSS, and of capillary leakage such as
leukocytosis with increasing VEGF levels. Ascitic ´¬‚uid obtained from the
study patients also con´¬ürmed high VEGF levels. These ´¬ündings suggest the
involvement of VEGF in the pathogenesis of capillary leakage in OHSS.
Lee et al. (1997) studied the relationship between serum and follicular ´¬‚uid
levels of VEGF, estradiol and progesterone in patients undergoing IVF, to
quantify the effects of hCG on serum levels of VEGF during early pregnancy
and to report serial measurements of serum and ascitic ´¬‚uid levels of VEGF in
a patient with severe OHSS. They found a signi´¬ücant ovarian contribution to
the circulating VEGF levels in early pregnancy. They concluded that elevated
serum VEGF levels may be a factor in the etiology of OHSS symptoms.
DÔÇ™Ambrogio et al. (1999) found that serum VEGF levels before starting
gonadotrophin treatment in women who have developed moderate forms of
OHSS showed no signi´¬ücant difference with a control group.
Aboulghar et al. (1999) observed higher VEGF plasma levels in patients
hospitalized for OHSS than in controls. The VEGF value dropped signi´¬ücantly
with clinical improvement, reaching minimal values after resolution. There was
a signi´¬ücant correlation between VEGF values and hematocrit as well as white
blood cell count.
VEGF AS A CAPILLARY PERMEABILITY AGENT IN OHSS
Agrawal et al. (1999) suggest that serum VEGF concentrations in IVF cycles
predict the risk of OHSS, and that VEGF increases may predict risk better than
the estradiol concentration, the number of follicles, and the number of oocytes,
which individually predict only 15%├Ç25% of cases.
In a prospective cohort study Enskog et al. (2001a) evaluated whether
differences in plasma VEGF165 concentrations exist during gonadotrophin
stimulation in IVF patients developing severe OHSS compared to matched
controls. They found that patients developing OHSS do not have raised plasma
VEGF165 levels during gonadotrophin stimulation. The lack of positive
correlation between VEGF165 levels and follicle numbers/progesterone in the
OHSS group suggests a disruption of the normal controlled follicular VEGF
expression in patients with OHSS.
The prognostic importance of serial cytokine changes in ascites and pleural
effusion in women with severe OHSS was evaluated and compared with ascitic
´¬‚uid in IVF cycles before oocyte retrieval. The results suggest that local cytokines
may be involved in the evolution of severe OHSS and possibly serve as
prognostic marker for this syndrome. Geva et al. (1999) concluded that pre-
ovulatory FF levels are not useful predictors for the development of OHSS. The
increased capillary permeability found in OHSS may be due to its systemic effect.
McElhinney et al. (2002) studied the variations in serum vascular
endothelial growth-factor-binding pro´¬üles and the development of OHSS,
and observed than patients who do not develop OHSS appear to have
a high-molecular-weight protein (a-2 macroglobulin) that binds VEGF to
a greater degree than occurs in patients who develop OHSS.
Gomez et al. (2002) reported that vascular endothelial growth-factor
receptor-2 activation induces vascular permeability in hyperstimulated rats
(Figures III.3 and III.4), and this effect is prevented by receptor blockade.
Fig. III.3: Time course of permeability among OHSS, control and PMSG groups,
at different time points after hCG
Reproduced with permission from Gomez et al. (2002). Endocrinology 143:4339├Ç48
60 PATHOPHYSIOLOGY OF OVARIAN HYPERSTIMULATION SYNDROME
Fig. III.4: Reverse transcriptase polymerase chain reaction detection of b-actin and whole
VEGF in the ovary and mesentery from OHSS, PMSG and control groups at different time
points after hCG
Reproduced with permission from Gomez et al. (2002). Endocrinology 143:4338├Ç49
VEGF AS A CAPILLARY PERMEABILITY AGENT IN OHSS
Mathur et al. (1967) studied whether serum VEGF levels can distinguish highly
responsive women who subsequently developed OHSS from women with a
similar ovarian response who do not. They found out that serum VEGF levels
are poorly predictive of subsequent OHSS in highly responsive women
undergoing assisted conception.
Artini et al. (1998) studied VEGF, interleukin-6 (lL-6) and interleukin-2
(lL-2) in serum and follicular ´¬‚uid of patients with OHSS. Patients presented
with follicular ´¬‚uid IL-6 levels higher than both the patients at risk and control
(p<0.05). On the day of the oocyte retrieval the patients developing OHSS
showed serum and follicular VEGF values higher than those of the patients
at risk (p<0.05). Serum and follicular ´¬‚uid IL-2 levels showed no differences
between the examined groups. IL-2, IL-6 and VEGF values were not correlated
with each other. The authors concluded that angiogenesis and in´¬‚ammation
processes are both present in severe OHSS.
Gomez et al. (2003) observed that administration of moderate and high
doses of gonadotrophins to female rats increases ovarian vascular endothelial
growth factor (VEGF) and VEGF receptor-2 expression that is associated with
Kitajima et al. (2004) observed that gonadotrophin-releasing hormone
agonist administration reduced VEGF, VEGF receptors, and vascular perme-
ability of the ovaries of hyperstimulated rats. They speculated that GnRH-a
treatment may prevent early OHSS by reducing vascular permeability through
the decrease in VEGF and its receptors.
VEGF in Follicular Fluid and GnRH Antagonist
Recent investigation of the impact of GnRH antagonist on IVF on the follicular
´¬‚uid VEGF content demonstrated no change. VEGF follicular ´¬‚uid content
is associated with embryo maturation, gonadotrophins dose and those of
follicular hypoxia. Some investigators reported an increase in follicular ´¬‚uid
VEGF concentration in poor responders, which is most likely a compensatory