| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Reproductive Endocrinology |
Medical Research Council Group in Fetal and Neonatal Health and Development, Lawson Research Institute, St. Joseph’s Health Center (D.J.H., E.A.), London, Ontario, Canada N6A 4V2; the Departments of Medicine (D.J.H., E.A.), Pediatrics (D.J.H.), and Physiology (D.J.H.), University of Western Ontario, London, Ontario, Canada N6A 5O5; the Institute of Experimental Clinical Research, Aarhus Kommunehospital (A.F.), DK-8000 Aarhus C., Denmark; and the Gynecological and Obstetrical Department, Skejby University Hospital (F.F.L., J.G.K.), DK-8200 Aarhus N., Denmark
Address all correspondence and requests for reprints to: Dr. D. J. Hill, Lawson Research Institute, St. Joseph’s Health Center, 268 Grosvenor Street, London, Ontario, Canada N6A 4V2. E-mail: dhill{at}lri.stjosephs.london.on.ca
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
A major complication of poorly controlled diabetic pregnancy is proliferative retinopathy. The angiogenic actions of FGF-2 both in vitro and in vivo (1) have prompted others to examine its presence in diabetic proliferative retinopathy outside of pregnancy. Both immunoreactive FGF-1 (acidic FGF) and FGF-2 (basic FGF) have been identified in pre-retinal membranes from patients with diabetic retinopathy and proliferative vitreoretinopathy and in the vitreous humor (10, 11, 12, 13, 14). Levels of FGF-2 were greatest in vitreous humor from patients with proliferative diabetic retinopathy in whom neovascularization of the disk or iris was evident (13). Such findings have implicated locally produced FGFs in the etiology of diabetic retinopathy. However, as pregnancy is normally associated with a circulating component of FGF-2, we have examined here whether this may differ between pregnant patients with type 1 diabetes with or without angiogenic eye disease.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Serum was collected, as described below, from 36 pregnant women with pregestational type 1 diabetes. Blood samples were drawn every 4 weeks between 14–30 weeks gestation and thereafter every 2 weeks until delivery. Serum was also collected 6 months postpartum. The project protocol was approved by the human ethics committees of Aarhus County (Aarhus, Denmark) and St. Joseph’s Health Center (London, Canada). Patients were categorized according to the White classification by the duration of diabetes and presence of complications (15). On this basis, 9 women were determined to be in White class B (debut age, >20 yr; duration, <10 yr; no late complications), 6 in White class C (debut age, 10–19 yr; duration, 10–19 yr; no late complications), 2 in White class Do (debut age, <10 yr; duration, >20 yr; no simplex retinopathy), 17 in White class D+ (simplex retinopathy), and 2 in White class R (proliferative retinopathy). All retinal examinations were performed by retinal photography with a dilated pupil by a blinded and experienced observer. Patients were then grouped into normal, simplex, and proliferative retinopathy states. No changes occurred in the severity of retinopathy in these patients during the course of pregnancy.
The mean age of the patients (±SD) was 29 ± 2 yr
(range, 22–36 yr). All patients were seen every second week in the
out-patient clinic from the beginning of pregnancy to 33 weeks and
thereafter weekly until term. All ambulatory check-ups were performed
by the same senior clinical physician, Dr. J. G. Klebe. All patients
performed home blood glucose measurements at least twice a week, with a
minimum of five measurements per day. At each clinic visit, body
weight, blood pressure, and daily blood glucose levels were determined,
and a 24-h urine sample was collected for the measurement of albumin,
glucose, and ketones. Blood was drawn for the measurement of
electrolytes, creatinine, and human placental lactogen (hPL).
Glycosylated hemoglobin (HbA1c) was measured every 4 weeks
up to 30 weeks and thereafter every second week until delivery.
HbA1c was determined by the method of Treveli et
al. (16) as modified by Schwartz et al. (17), using
Bio-Rex 70 (Bio-Rad Laboratories, Munich, Germany) as an ion exchange
resin. The normal range was 4.4–6.4%, with an intraassay coefficient
of variation of less than 1% and an interassay coefficient of
variation of less than 3%. 
-Fetoprotein was measured once before 24
weeks gestation. All patients were interviewed during gestation and
after delivery with regard to diet. Although patients increased their
overall food consumption over time, no patient had an altered dietary
composition.
Ambulatory 24-h blood pressure measurements were performed in all subjects three times during pregnancy and once postpartum. The blood pressure measurements were recorded each hour by means of a portable blood pressure monitor (Spacelab 90207, Redmond, WA), and a mean 24-h blood pressure was calculated as previously described (18). Ultrasound scan was performed twice early in pregnancy and every sixth week. Fetal weight estimations were performed at 32 and 36 weeks. Funduscopic examination was performed at least three times during pregnancy. Normal glycemia was maintained by insulin therapy together with a standard diabetes diet with high carbohydrate content. The majority of patients were delivered after rupture of membranes at the beginning of the 38th week. Earlier inductions were instituted if complications of pregnancy occurred. Newborn weight and crown-heel length were recorded for each infant. Body weight percentile values were determined from the standard charts of Keen and Pearse (19). Blood (10 mL) was allowed to clot at 4 C and was centrifuged at 1000 x g, and serum was collected and stored at -20 C until analysis.
Reagents
Recombinant human (h) FGF-2 was purchased from Upstate Biotechnology (Lake Placid, NY), and a rabbit polyclonal antibody (Ab 773) raised against the 1–24 synthetic fragment of bovine FGF-2 was provided by Dr. A. Baird (PRIZM Pharmaceuticals, San Diego, CA). The antibody demonstrates less than 1% cross-reactivity with FGF-1, FGF-3 (int-2), FGF-4 (hst/ks), FGF-5, FGF-6, and FGF-7. The cross-reactivity with FGF-8 and FGF-9 is not known. However, each of these peptides shows low sequence homology with the peptide fragment of FGF-2 used as the antigen (20). The same antibody was found to block the mitogenic actions of exogenous or endogenous FGF-2 on ovine growth plate chondrocytes in vitro (21). The goat anti-rabbit IgG used within the FGF-2 RIA was obtained from Interscience (Markham, Canada). Heparin-Sepharose CL-6B was purchased from Pharmacia LKB Biotechnology (Uppsala, Sweden), and 125I-labeled FGF-2 (SA, 970 Ci/mmol) was purchased from Amersham International (Mississauga, Canada).
Sample preparation and FGF-2 RIA
The FGF-2 immunoreactivity present in pre- and postpartum
maternal serum (1 mL) was extracted using heparin-Sepharose affinity
chromatography as described in detail previously (6). Pregnancy serum
was seeded with either 125I-labeled FGF-2 or with 3 nmol/L
unlabeled hFGF-2 before heparin-Sepharose extraction and RIA to
determine the efficiency of extraction. Both analyses showed a recovery
of between 75–85% of added FGF-2. RIA for FGF-2 was performed, as
previously described (6), in polypropylene assay tubes with a reaction
mixture consisting of 300 mL assay buffer [0.01 mol/L phosphate buffer
containing 0.1% (wt/vol) sodium azide, 0.03 mol/L disodium
ethylenediamine-tetraacetic acid, 0.15 mol/L NaCl, and 0.3% (wt/vol)
BSA, pH 7.4], 100 mL diluted primary antiserum (Ab 773; final
concentration, 1:50,000), and 100 mL test sample or recombinant hFGF
standard preparation. A standard curve ranging from 5.5–550 fmol/tube
was used. After mixing, the reagents were incubated for 24 h at 4
C before the addition of 100 µL 125I-labeled FGF-2
(
20,000 cpm/tube). After an additional incubation of 24 h at 4
C, 100 mL goat antirabbit IgG [125 U diluted to 25 mL with assay
buffer and containing 1% (vol/vol) normal rabbit serum] were added to
each tube and mixed, and assay tubes were allowed to stand at 4 C for
24 h before the addition of 1 mL assay buffer. Assay tubes were
centrifuged at 2,500 x g for 1.5 h to separate
bound from free radiolabel. The supernatant was aspirated, and the
bound fraction was counted by 
-spectroscopy. Nonspecific binding was
approximately 5% of the added counts; the within-assay coefficient of
variation was 5%, and the between-assay coefficient of variation was
12%. The minimum level of detection was 20 fmol/tube FGF-2. Each
sample was assayed at up to four serial dilutions, and results were
expressed as picomoles per L FGF-2 related to the original volume of
the sample, where 1 pmol/L equals 18 ng/L.
Measurement of hPL, insulin, creatinine, blood glucose, and albumin
hPL, HbA1c, and serum and urinary creatinine were measured using standard techniques. Blood glucose was measured by the glucose oxidase method using a Yellow Springs glucose analyzer (Yellow Springs, OH). A linear measurement of blood glucose values was possible up to 25 mmol/L. Urinary albumin excretion was measured in urine samples collected over 24 h before each ambulatory visit. The urinary albumin concentration was measured by RIA using polyethylene glycol separation, as previously described (22). Intra- and interassay coefficients of variation were less than 5% and 10% for all assays. Serum insulin levels were determined by a double antibody RIA obtained from Kabi Pharmacia Diagnostics (Uppsala, Sweden). The standard preparation was recombinant human insulin, and the minimum level of detection was less than 2 µU/mL. The intra- and interassay coefficients of variation were less than 6% and 7%, respectively.
Statistical analysis
Differences between mean serum immunoreactive FGF-2 levels; between mean values for serum insulin, glucose, creatinine, albuminuria, HbA1c, and blood pressure; and between parameters of fetal size were assessed in different patient groups using ANOVA. Post-hoc analysis was performed using the Bonferroni/Dunn test with a significance level of 5%. Correlations between the immunoreactive FGF-2 content of maternal serum and other parameters were analyzed by linear regression analysis.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
), which were significantly greater than
those in patients without retinopathy at 18 and 22 weeks gestation.
However, HbA1c has been reported to be a relatively poor
predictor of average blood glucose concentrations in pregnant diabetic
woman (23). When mean weekly blood glucose values were compared between
patients with or without retinopathy, they were significantly elevated
in the former group only at 32 weeks gestation (Fig. 1).
Although the mean duration from the onset of diabetes differed
substantially between patients according to White classification, no
significant differences were seen in either mean birth weight or length
of the offspring between patient groups (Table 2). The
mean birth weight percentile was greater than 70% for gestational age
in all groups, as reported previously for infants of pregestational
diabetic mothers by us and others (9, 24).
|
|
|
. FGF-2 was detectable in serum from all patients at 14
weeks gestation, and the highest levels were seen between 22–26 weeks,
as previously reported by us (6). Thereafter, circulating levels of
FGF-2 declined toward term and were extremely low, but still
detectable, in most patients postpartum. No significant differences
were seen between serum FGF-2 levels in women of White classifications
B, C, and D0 at any gestational age. However, when patients
with retinopathy (D+ and R) were compared with those
without, the mean FGF-2 levels were significantly greater in the former
between 22–32 weeks gestation. In pregnant women with retinopathy,
serum FGF-2 levels of 400–500 pmol/L were found in the second
trimester. From 32 weeks gestation until term, there was no longer a
significant difference in serum FGF-2 between patients with or without
retinopathy. A significant positive correlation existed between serum
FGF-2 levels and HbA1c values at 22 (r2 = 0.15;
P < 0.05), 30 (r2 = 0.25;
P < 0.02), and 34 (r2 = 0.87;
P < 0.001) weeks gestation. No significant
correlations existed between serum FGF-2 and weekly mean blood glucose
values.
|
). Creatinine clearance fell with gestational age, but
did not significantly differ between patients with or without
retinopathy, although there was a trend for creatinine clearance to be
lower in the former (Fig. 3c). Similarly, serum creatinine did not
differ between patient groups, nor did systolic or diastolic blood
pressure, as measured during clinic visits (Fig. 3b), which tended to
increase with the progression of pregnancy in all patients. Mean
(±SEM) diastolic 24-h blood pressure measurements obtained
at home by patients with retinopathy were: at 14–18 weeks, 74 ±
2 mm; at 22–26 weeks, 72 ± 2 mm; and at 32–34 weeks, 79 ±
2 mm Hg. These did not significantly differ from the mean values in
patients without retinopathy (69 ± 2, 69 ± 2, and 74
± 1 mm Hg, respectively). No significant nighttime dip in blood
pressure was seen in any patient, and no changes in diurnal blood
pressure were seen between patients with and without retinopathy.
Therefore, there was no indication of serious impairment of renal
function or of hypertension in the pregnant women with retinopathy. As
it is likely that some or all of the circulating FGF-2 derived from the
placenta, we measured circulating hPL as an indicator of placental
endocrine function (Fig. 3d). Levels of hPL rose progressively with
gestational age, but did not differ in patients with or without
retinopathy other than at 30 weeks, when significantly greater values
were seen in women with retinopathy. No significant correlations were
seen at any gestational age between serum hPL and FGF-2, or between
FGF-2 and systolic blood pressure or creatinine clearance rate.
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
A likely source of FGF-2 in the maternal circulation is the placenta; this is supported by a substantial fall in circulating FGF-2 in most patients postpartum (6). Placenta has been reported to express FGF-2 messenger ribonucleic acid and contain immunoreactive FGF-2 peptide (8, 25). We recently showed that placentas from pregnant women with type 1 diabetes contain a greater abundance of FGF-2 and FGFR1 messenger ribonucleic acids than placentas from control pregnancies (26). In placentas from infants born prematurely, FGF-2 detected by immunocytochemistry is abundant in the syncytiotrophoblast, but much less intense staining is seen in placentas from normal pregnancies at term. Placentas from women with type 1 diabetes retained intense staining for FGF-2 in the syncytiotrophoblast at term. It has not been determined whether the expression of FGF-2 is greater in placentas from diabetic women also exhibiting retinopathy. Circulating hPL values were measured as a general indicator of placental endocrine function. These did not differ substantially in diabetic women with or without retinopathy, other than a marginal increase in the former at 30 weeks gestation, suggesting that if the circulating FGF-2 derives mainly from placenta, the greater levels seen with retinopathy are a specific association and not related to a general increase in placental protein production. Mean birth weight or length did not differ between diabetic pregnancies with or without retinopathy. Although placental weights were not recorded, it is unlikely that mean placental size would differ substantially between the two groups. Six months after delivery, most patients in this study had low (>100 pmol/L), but measurable, levels of FGF-2 remaining in their circulations. We have previously shown that nonpregnant women without diabetes have no detectable serum FGF-2 (6), suggesting that the type 1 diabetes alone may be associated with the appearance of FGF-2 in blood.
Assessment of clearance is complicated by the presence of a circulating FGF-binding protein (FGF-BP) in pregnancy, derived from the extracellular domain of FGFR1 (6). It is not known whether the FGF:FGF-BP complex extends the biological half-life of the circulating FGF-2 or if it may protect FGF-2 from proteolysis in the extracellular fluids. Although FGF-BP can be visualized by Western blot in pregnancy serum (6), at present we have no quantitative method to determine whether the FGF-2/FGF-BP ratio is substantially altered in diabetic pregnancies with or without retinopathy. The diabetic patients with retinopathy examined in this study did not suffer from macroalbuminuria, significantly altered creatinine clearance, or relative hypertension. Therefore, there is no indication that the elevated serum FGF-2 levels seen in those with retinopathy resulted from altered renal function and protein clearance.
We previously showed, using cross-sectional measurements, that during normal pregnancy, circulating maternal FGF-2 values were greatest between 28–31 weeks gestation (6). The present study, using longitudinal analysis, confirmed that a peak of circulating FGF-2 occurs in midgestation, but this is seen at 22–26 weeks in diabetic pregnancy. Without comparable sampling from normal pregnancies we cannot be sure that this reflects a physiological difference. However, as the greatest mean value for serum FGF-2 was seen at 18 weeks gestation in diabetic women with retinopathy (480 ± 102 pmol/L), whereas in normal pregnancies the mean value was less than 100 pmol/L at this time (6), it seems likely that the normal ontogeny of serum FGF-2 is substantially altered in type 1 diabetes. This may be directly related to the metabolic control of the diabetic patient. When sampled at term, maternal FGF-2 was positively correlated with HbA1c values in pregnancies complicated by diabetes (9). In the present study, HbA1c levels were significantly greater at 18 and 22 weeks gestation in diabetic pregnancies complicated by retinopathy than in those without, corresponding to significantly greater serum FGF-2 levels also seen at this stage of pregnancy. Serum FGF-2 levels showed significant positive correlation with HbA1c levels at 18, 30, and 34 weeks. However, the elevated FGF-2 in patients with retinopathy is unlikely to be linked directly to blood glucose levels because mean weekly blood glucose values did not differ between patients with or without retinopathy other than at 32 weeks gestation, which is after differences in circulating FGF-2 were found. The altered circulating FGF-2 may relate to the anatomical changes and infarctions reported to occur in the placentas of pregnant diabetics if placenta is a major source of serum FGF-2, as injury elsewhere in the body is associated with increased expression of FGF-2 (1, 27).
The amount of FGF-2 found in serum during the early second trimester in diabetic patients with retinopathy would be sufficient to cause endothelial cell proliferation in vitro (28). Whether the circulating FGF-2 is causally related to retinopathy in diabetic pregnancy is unknown, but evidence exists to suggest that this is possible. Immunoreactive FGF-2 is present in the vitreous humor of diabetic patients and in preretinal membranes, and is greatest in patients with proliferative disease (11, 13, 15). FGF-BPs representing truncated FGFR1 receptor have also been identified in vitreous humor (29). Given the proven angiogenic capacity of FGF-2 in the eye in animal models (1), it seems likely that a local production of FGF-2 could contribute to proliferative retinopathy. In diabetic pregnancy, locally produced FGF-2 may be supplemented by that derived from the circulation, although the pharmacokinetics in relation to circulating FGF-BP are not known.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received November 1, 1996.
Revised January 14, 1997.
Accepted January 30, 1997.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  F. F. Lauszus, J. G. Klebe, T. Bek, and A. Flyvbjerg Increased Serum IGF-I During Pregnancy Is Associated With Progression of Diabetic Retinopathy Diabetes, March 1, 2003; 52(3): 852 - 856. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Kurz, L. Hefler, H. Zeisler, C. Schatten, P. Husslein, and C. Tempfer Maternal Basic Fibroblast Growth Factor Serum Levels Are Associated With Pregnancy-Induced Hypertension Reproductive Sciences, January 1, 2001; 8(1): 24 - 26. [Abstract] [PDF]
|
|
|
|
 |
|
 A. Eliakim, Y. Oh, and D. M. Cooper Effect of single wrist exercise on fibroblast growth factor-2, insulin-like growth factor, and growth hormone Am J Physiol Regulatory Integrative Comp Physiol, August 1, 2000; 279(2): R548 - R553. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
) and those without
(B, C, and Do; n = 17; 
). Figures show the
mean ± SEM. *, P < 0.05.
