0021-972X/99/$03.00/0 Journal of Clinical Endocrinology and Metabolism Copyright © 1999 by The Endocrine Society
Vol. 84, No. 2 Printed in U.S.A.
Reduced Retinal Vascularization in Children with Growth Hormone Deficiency* ¨ M, ELISABETH SVENSSON, BJO ¨ RN CARLSSON ANN HELLSTRO AIMON NIKLASSON, AND KERSTIN ALBERTSSON-WIKLAND Department of Clinical Neuroscience’s, Section of Ophthalmology, International Pediatric Growth Research Center, Department of Pediatrics, Research Center for Endocrinology and Metabolism, Department of Internal Medicine, University of Goteborg; and the Biostatistics Branch, Department of Mathematical Statistics, Chalmers University of Technology, Go¨teborg, Sweden ABSTRACT The neovascularization in diabetic retinopathy is believed to involve locally produced angiogenic factors. In addition, there are indications that GH may influence retinal vascularization. To investigate the role of GH in retinal vascularization, we examined the retinal vascular pattern in children with congenital GH deficiency. Retinal vessel morphology was evaluated by digital image analysis of ocular fundus photographs in 39 children (5 girls and 34 boys, aged 3.6 –18.7 yr) with congenital GH deficiency, and it was compared to that of 100 healthy controls. Twenty children had received GH treatment (0.1
IU/kg daily). All children were born at term, and none of the children had any clinical signs of ocular disease or reduced vision. Children with GH insufficiencies, regardless of whether they were treated with GH, had a significantly lower number of vascular branching points than the reference group (P , 0.0001). Thirty-three percent of the GH-insufficient individuals had a number of vascular branching points less than or equal to the fifth percentile of the reference group. The reduced retinal vascularization observed in children with congenital GH deficiency suggests that GH may be of importance for angiogenesis. (J Clin Endocrinol Metab 84: 795–798, 1999)
S
EVERAL decades ago it was reported that pituitary ablation resulted in remission of diabetic retinopathy (1–3) and that the remission was related to the reduction in serum levels of GH (1). However, the specific role of GH in neovascularization has been controversial and much attention has lately instead focused on locally produced angiogenic factors such as vascular endothelial growth factor, fibroblast growth factor, and insulin-like growth factors I and II (IGF-I and IGF-II) (4 – 6). Experiments in rodents have provided new evidence that the GH/IGF-I axis may influence angiogenesis in the brain (7) and the development of retinal neovascularization through interaction with locally produced factors (8). Interestingly, when examining the retinal vascular pattern in children with congenital GH deficiency, we found that this group of children has reduced retinal vascularization.
included. Thirty-three of these children had isolated GH deficiency, and 6 children had multiple pituitary deficiencies. The median height at the time of the GH investigation was 23.0 sd score (range, 29.4 to 20.8 sd score). Twenty of these children had their fundus photographed (median, 7.8 yr; range, 2.4 –11 yr) after GH treatment was started. One hundred healthy individuals (56 boys and 44 girls) born at term (age range, 2.6 –19.6 yr) constituted a reference group for evaluation of ocular fundus morphology. Detailed data for these children and adolescents were presented previously (10). The study was approved by the ethics committee at the Medical Faculty, Goteborg University. Informed consent was obtained from the parents and, if they were old enough, from the children themselves. All children had an eye examination, including fundus photography. Visual acuity ranged from 20/40 to 20/20 (median, 20/20). Refraction ranged from 21.5 to 12 diopters. All ocular fundus photographs were analyzed quantitatively, using digital image analysis (11). The retinal vascularization was analyzed with respect to the number of branching points and the tortuosity of arteries and veins. Measurements of retinal arterioles and venules (referred to as arteries and veins) were made by tracing each vessel (path length) from its origin on the optic disc to a reference circle with a radius of 3.0 mm from the geometric center of the optic disc. The index of tortuosity for arteries and veins was defined as the path length of the vessel divided by the linear distance from the vessel origin to the reference circle. Vessels were also marked from their branching point to the reference circle, and the total number of branching points (arteries and veins), i.e. the number of retinal vessels within this area was calculated. The maximal GH level was estimated in each child, before the start of treatment with GH, from values obtained during the measurement of spontaneous 24-h GH secretion (12, 13) (n 5 23) and/or during an arginine insulin tolerance test (12) (n 5 36). GH deficiency was defined as a maximum GH level below 20 mU/L, as measured by polyclonal antibody-based immunoradiometric assay (Pharmacia & Upjohn, Inc., Uppsala, Sweden) with WHO International Reference Preparation 66/ 127 as standard. IGF-I was measured as previously described (14) in 36 children (Table 1). Bone age was assessed according to the method of Tanner and Whitehouse (15). The mean of the two eye measurements represented the value of each
Subjects and Methods From 1992–1994, all children referred to the Children’s Hospital (Go¨teborg, Sweden) for impaired growth [body height below 22 sd compared with the Swedish reference values (9)] with suspicion of congenital GH deficiency had an ophthalmological examination (n 5 30). In addition, 19 children treated at our unit, representing those with the most severe GH deficiency, had an ophthalmological examination between 1995 and 1996. Seven children were excluded due to preterm birth, and three were excluded due to a combination of other diagnoses. Thus, 39 full-term children (5 girls and 34 boys, aged 3.6 –18.7 yr) were Received May 7, 1998. Revision received August 24, 1998. Accepted November 9, 1998. Address all correspondence and requests for reprints to: Ann Hellstro¨m, M.D., Ph.D., Section of Pediatric Ophthalmology, Sahlgrenska ¨ stra, S-416 85 Go¨teborg, Sweden. E-mail: University Hospital/O
[email protected]. * This work was supported by grants from The Swedish Medical Research Council (7509, 10863, 11331, and 11576).
795
¨ M ET AL. HELLSTRO C-age, Chronological age; B-age, bone age [Tanner-Whitehouse stage II, 20 bones (15)]; AITT, arginine insulin tolerance test; BP-3, IGF-binding protein-3. GH deficiency was defined as maximum GH (GH max) below 20 mU/L.
14 0.75 22.67 1.87 14 1.45 22.99 3.26 17 22.15 24.35 0.09 17 22.44 26.62 0.18 19 5.9 0.04 17.6 19 2.7 0.9 7.5 19 21.32 23.83 20.11 17 11.64 5.78 20.5 10 13.66 3.73 20.21 19 9.1 3.2 13.9 19 11.1 5.9 17.4 n Median Minimum Maximum
SD
IGF-I
19 20.76 27.68 0.05
20 3.3 1.0 10.7
13 7.01 1.05 18.68
BP-3 SD score IGF-I SD score
Values obtained after diagnosis
BP-3 SD score IGF-I SD score AITT GH max
Values obtained at diagnosis
24-h GH max B-age (yr) C-age (yr) BP-3 SD score AITT GH max
Values obtained at diagnosis
24-h GH max B-age (yr) C-age (yr)
Fundus photo taken before GH treatment (n 5 19)
TABLE 1. Characteristics of GH axis in 39 children with congenital GH deficiency
Fundus photo taken during GH treatment (n 5 20)
796
JCE & M • 1999 Vol 84 • No 2
FIG. 1. Number of retinal vascular branching points and index of tortuosity for arteries and veins in children with congenital GH deficiency. Ocular fundus photo taken before GH (circles; n 5 19) and during (range, 2.4 –11 yr; triangles; n 5 20) GH treatment. The shaded area depicts the 5th to the 95th percentile range, and the center line indicates the median for the healthy reference group (10). fundus variable of an individual, and the data were analyzed using the sign test. Probability was calculated using the modified Mann-Whitney test formula (16).
Results
Children with congenital GH deficiency had a significantly reduced number of vascular branching points compared to a reference group of healthy children (10) (median number, 19.5 and 23, respectively; P , 0.0001; Fig. 1). Thirtythree percent of the GH-insufficient individuals had fewer or the same number of vascular branching points compared to the fifth percentile of the reference group. The probability for a randomly selected individual with congenital GH deficiency to have a lower number of vascular branching points than a randomly selected individual in the reference group was 75%. There was no difference regarding the tortuosity index for arteries (median, 1.08 and 1.10 for children with congenital GH deficiency and controls, respectively) or veins (median, 1.07 for both groups) between the two groups. The pattern of retinal vascularization in children with congenital GH deficiency was independent of whether they had received GH treatment (median number of vascular branch-
REDUCED RETINAL VASCULARIZATION IN GH DEFICIENCY
ing points, 19 and 20, respectively; tortuosity for arteries, 1.08 and 1.09, respectively; tortuosity for veins, 1.07 for both groups; Fig. 1). Discussion
The possible role of GH in human angiogenesis and retinal neovascularization is controversial (17). However, the observation that pituitary ablation had beneficial effects on diabetic retinopathy has provided circumstantial evidence for a role for GH in the pathophysiology of retinal neovascularization (1–3). In this study we showed that children with congenital GH deficiency have reduced retinal vascularization (Fig. 2). Our study indicates that GH may participate in the physiological regulation of angiogenesis in humans. This is in line with recent studies in the rat that indicated that GH and IGF-I may be of physiological importance for the regulation of cerebral microvasculature (7). Retinal vascularization was also reduced in the children with congenital GH deficiency that had received GH treatment for several years. Vascularization of the retina normally occurs during fetal development, with little or no vascularization after birth (18, 19). Thus, it is possible that GH acts as a permissive factor for other angiogenic stimuli during fetal development. Previous studies in rodents support the hypothesis that GH may act as a permissive hormone, as both vascular remodeling in response to hypertension (20) and ischemia-induced retinal vascularization (8) are inhibited in the absence of circulating GH. However, an increased circulating level of GH is not sufficient to induce these effects in the absence of other angiogenic stimuli (8). The mechanism by which GH influences angiogenesis is not known. It is possible that GH directly influences this process, as GH receptors are expressed in human blood vessels in the human fetus (21) and GH stimulates the proliferation of human microvascular cells in vitro (22). Alternatively, GH may exert its effects through circulating or locally produced IGF-I. In fact, patients with severe retinopathy have increased levels of IGF-I in the vitreous (23). Further-
797
more, there are several indications that IGF-I may influence angiogenesis (5). Although maintenance of the vasculature and angiogenesis appear to be regulated to a large extent by locally produced growth factors in response to stimuli such as hypoxia, there are several studies indicating that circulating hormones and factors other than GH may influence the production or action of locally produced factors (6, 8). For example, angiostatin is a recently identified circulating inhibitor of angiogenesis (6), and IGF-I and erythropoietin have angiogenic properties (5, 24). From a physiological perspective, locally produced factors may respond to local signs of an increased vascular demand such as hypoxia, whereas circulating factors may have a permissive role and allow angiogenesis when appropriate. For example, circulating factors may allow a locally produced signal that stimulates angiogenesis when the nutritional status of the organism is sufficient. An increased understanding of the interaction between locally produced factors and the GH/IGF-I system may open new therapeutical possibilities in the treatment and prevention of vascular disease. The possible effect of GH treatment on retinal vascularization is controversial. The effect of GH treatment on the number of vascular branching points in this study should be interpreted with caution, as only half of the study group received GH treatment and was compared to children that did receive GH treatment. Furthermore, the GH treatment was introduced at different ages and was of varying duration. With these limitations in mind, the present study did not indicate any correlation between the length of GH treatment and the number of vascular branching points. However, a recent study identified two subjects with retinopathy associated with GH therapy (25). Clearly, more studies on the possible role of GH treatment on retinal vascularization in children and adults are warranted. We conclude that children with congenital GH deficiency, regardless of treatment with GH, have reduced retinal vascularization. The present study extends recent studies in
FIG. 2. Reduced retinal vascularization in a 7-yr-old boy with congenital GH deficiency (right) and normal vascularization in a healthy 6-yr-old boy (left).
¨ M ET AL. HELLSTRO
798
rodents that indicate that GH may act as a permissive hormone for angiogenesis.
13.
Acknowledgments The authors are grateful for the cooperation of Yahua Chen, M.D., and for the practical help with patients given by Ward 34T, Cecilia Axelson, and the staff at the Department of Ophthalmology.
14.
15.
References 1. Wright AD, Kohner EM, Oakley NW, Hartog M, Joplin GF, Russel Frazer T. 1969 Serum growth hormone levels and the response of diabetic retinopathy to pituitary ablation. Br Med J. 2:346 –348. 2. Lundbaek K, Christensen NJ, Jensen VA. 1970 Diabetes, diabetic angiopathy and growth hormone. Lancet. 2:131–133. 3. Teuscher A, Eschen F, Konig H, Zahnd G. 1970 Long term effects of transsphenoidal hypophysectomy on growth hormone, renal function and eye ground in patients with diabetic retinopathy. Diabetes. 19:502–518. 4. Battegay EJ. 1995 Angiogenesis: mechanistic insights, neovascular diseases, and therapeutic prospects. J Mol Med. 73:333–346. 5. Delafontaine P. 1995 Insulin-like growth factor I and its binding proteins in the cardiovascular system. Cardiovasc Res. 30:825– 834. 6. Folkman J. 1995 Clinical applications of research on angiogenesis. N Engl J Med. 333:1757–1763. 7. Sonntag WE, Lynch CD, Cooney PT, Hutchins PM. 1997 Decreases in cerebral microvasculature with age are associated with the decline in growth hormone and insulin-like growth factor 1. Endocrinology. 138:3515–3520. 8. Smith LEH, Kopchick JJ, Chen W. 1997 Essential role of growth hormone in ischemia-induced retinal neovascularisation. Science. 276:1706 –1709. 9. Karlberg J. 1989 A biologically-oriented mathematical model (ICP) for human growth. Acta Paediatr. 350:70 –94. 10. Hellstro¨m A, Svensson E. 1998 Reference intervals of optic disc and retinal vessel variables in healthy children. Acta Ophthalmol. 76:260 –267. 11. Stro¨mland K, Hellstro¨m A, Gustavsson T. 1995 Morphometry of the optic nerve and retinal vessels by computer-assisted image analysis. Graefes Arch Clin Exp Ophthalmol. 233:150 –153. 12. Albertsson-Wikland K, Rosberg S, Karlberg J, Groth T. 1994 Analysis of
16. 17. 18. 19. 20.
21.
22.
23.
24. 25.
JCE & M • 1999 Vol 84 • No 2
24-hour growth hormone profiles in healthy boys and girls of normal stature: relation to puberty. J Clin Endocrinol Metab. 78:1195–1201. Boguszewski M, Rosberg S, Albertsson-Wikland K. 1995 Spontaneous 24hour growth hormone profiles in prepubertal small for gestational age children. J Clin Endocrinol Metab. 80:2599 –2606. Boguszewski M, Jansson C, Rosberg S, Albertsson-Wikland K. 1996 Changes in serum insulin-like growth factor (IGF-I) and IGF-binding protein-3 levels during growth hormone treatment in prepubertal short children born small for gestational age. J Clin Endocrinol Metab. 81:3902–3908. Tanner JM, Whitehouse R, Marshall W, Healey M, Goldstein H. 1975 Assessment of skeletal maturing and prediction of adult height (TW2 method). London: Academic Press. Altman DG. 1994 Practical statistics for medical research. London: Chapman & Hall. 186. Sharp PS. 1995 The role of growth factors in the development of diabetic retinopathy. Metabolism. 44:S72–75. Ashton N. 1970 Retinal angiogenesis in the human embryo. Br Med Bull. 26:103–106. Roth AM. 1977 Retinal vascular development in preterm infants. Am J Ophthalmol. 84:636 – 640. Fokow B, Isaksson OGP, Karlstro¨m G, Lever AF, Nordlander M. 1988 The importance of hypophyseal hormones for structural cardiovascular adaptation in hypertension. J Hypertens. 6(Suppl):S166 –S169. Werther GA, Haynes K, Waters MJ. 1993 Growth hormone receptors are expressed on human fetal mesenchymal tissues-identification of messenger ribonucleic acid and GH-binding protein. J Clin Endocrinol Metab. 76:1638 –1646. Rymaszewski Z, Cohen RM, Chomczynski P. 1991 Human growth hormone stimulates proliferation of human retinal microvascular cells in vitro. Proc Natl Acad Sci USA. 88:617– 622. Meyer-Schwickerath R, Pfeiffer A, Blum WF. 1993 Vitreous levels of insulinlike growth factors I and II and the insulin-like growth factor binding proteins 2 and 3 increase in neovascular eye disease. J Clin Invest. 92:2620 –2625. Anagnostou A, Liu Z, Steiner M. 1994 Erythropoietin receptor mRNA expression in human endothelial cells. Proc Natl Acad Sci USA. 91:3974 –3978. Koller EA, Green L, Gertner JM, Bost M, Malozowski N. 1998 Retinal changes mimicking diabetic retinopathy in two nondiabetic, growth hormone-treated patients. J Clin Endocrinol Metab. 83:2380 –2383.
