- Email: [email protected]
Nonsuppressible insulin-like activity and somatomedin C levels in normal pregnant women, in pregnant women with gestational diabetes, and in umbilical cord blood of mature and premature infants
Nonsuppressible insulin-like activity and somatomedin C levels in normal pregnant women, in pregnant women with gestational diabetes, and in umbilical cord blood of mature and premature infants Naguib A. Samaan, M.D., Ph.D., Rena Vassilopoulou-Sellin, M.D., Pamela N. Schultz, B.S., R.N., Manuel E. Rivera, M.D., and Berel Held, M.D. Houston, Texas This study was undertaken to determine the significance of the changes in nonsuppressible insulin-like activity as measured by the fat pad assay and by the levels of immunoreactive somatomedin C, growth hormone, and human placental lactogen in sera of term normal pregnant women, mothers who delivered prematurely, and women with gestational diabetes at term as compared to normal nonpregnant subjects. These hormones were also measured in the umbilical cord blood of these patients at the time of delivery to determine the possible mechanisms of the fetal growth in utero. Our investigations showed that (1) nonsuppressible insulin-like activity is elevated during pregnancy, but its level was lower in mothers with gestational diabetes in spite of significantly higher serum human placental lactogen compared with normal pregnant mothers; (2) nonsuppressible insulin-like activity is significantly lower in premature infants than in term infants; (3) somatomedin C levels were significantly elevated in pregnant mothers in spite of suppression of growth hormone; (4) nonsuppressible insulin-like activity and somatomedin C levels in infants of mothers with gestational diabetes were not significantly elevated in spite of higher birth weight, indicating that nonsuppressible insulin-like activity and somatomedins are not the only factors responsible for the increase of birth weight of children of diabetic mothers; (5) there was marked discordance between the growth hormone level in the neonates and somatomedin C levels. (AM J OBSTET GYNECOL 1985;153:457-61.)
Key words: Nonsuppressible insulin-like activity, somatomedins, mother, fetus
The mechanism responsible for fetal growth in utero is poorly understood. The isolation and the description of the sequence of the insulin-like growth factors I and II by Rinderknecht and Humbel in 1976 1 shed some light on this important subject. The isolation of these growth factors, also known as somatomedins, was preceded by the description of "typical" and "atypical" insulin-like activity by Samaan et al., 2 later called "suppressible" and "nonsuppressible" by Froesch et al., 3 and of a serum component (sulfation factor), mediating the growth hormone-induced incorporation of sulfate into cartilage by Salmon and Daughaday. 4 Numerous atFrom the Section of Endocrinology, Department of Medicine, The University of Texas M. D. Anderson Hospital and Tumor Institute at Houston, and the Department of Obstetrics and Gynecology, The University of Texas Medical School at Houston. This study was supported by Grant No. CA-35040 awarded by the National Cancer Institute and the Nancy Carmichael Gift Fund. Presented at the Thirty-second Annual Meeting of the Society for Gynecologic Investigation, Phoenix, Arizona, March 20-23, 1985. Reprint requests: Naguib A. Samaan, M.D., Ph.D., Section of Endocrinology, The University of Texas M. D. Anderson Hospital and Tumor Institute at Houston, 6723 Bertner Ave., Houston, TX 77030.
tempts have been made to identify these substances, their site of formation, and interrelationships. Samaan et al. 2 • 5 · 7 produced strong evidence that nonsuppressible insulin-like activity, or "atypical" insulin-like activity, was formed in the liver after regular insulin, "typical insulin-like activity," or insulin extracted from the pancreas was infused into the liver. Recently Daughaday et al. 8 also showed evidence that the liver plays a major part in formation of sulfation factor when growth hormone as well as insulin was infused into the liver, and they introduced the name somatomedin for the sulfation factor. Fryklund et al. 9 succeeded in purifying somatomedin A and Underwood et al. 10 purified and separated somatomedin C from adult human plasma. Rinderknecht and Humbel, 1 in addition to purifying insulin-like growth factors I and II, also determined their amino acid sequence, which was found to be similar to proinsulin. Zapf et al. 11 found that both insulin-like growth factors I and II fulfill all the criteria of the somatomedins. lnsulinlike growth factor I was reported to be structurally similar to somatomedin C, whereas insulin-like growth
457
458 Samaan et al.
October 15, 1985 Am J Obstet Gynecol
factor II represents a fraction of the nonsuppressible insulin-like activity. 11 Insulin-like growth factor II was found to be more potent than insulin-like growth factor I on adipose tissue but less potent in cartilage. 11 Somatomedin A and somatomedin C also were found to share the same biologic properties but to differ in their amino acid compositions. 11 Insulin-like growth factor I and II or the somatomedins have been measured in the sera of pregnant women and the umbilical cord by various techniques ~ ; however, the correlation between the somatomedins and nonsuppressible insulinlike activity as well as the hormones of pregnancy in mother and fetus need further clarification. This study was undertaken (1) to measure the maternal and fetal levels of somatomedin C and nonsuppressible insulin-like activity in relation to selected hormone levels at term and (2) to relate these biochemical findings with the clinical settings, in particular diabetes and premature birth. 12 16
Material and methods
Sera were collected by venipuncture at the time of delivery from 17 normal women at term (group A, gestation of 38 to 41 weeks), 15 normal women who delivered prematurely (group B, gestation of 26 to 36 weeks), and 14 women with gestational diabetes who did not receive insulin treatment but were treated with diet alone (group C, gestation of 37-41 weeks). Peripheral serum samples also were collected from 15 normal nonpregnant women matched for age (group D). Umbilical cord blood was obtained from groups A, B, and C at delivery immediately after the cord was clamped. The blood was collected in nonheparinized tubes from the placental side, which contained both arterial and venous blood. The serum was·separated and kept at - 20° C until the time of assay. This study was done with the approval of the Human Protection Committee, and informed consent was obtained from each mother. Immunoreactive growth hormone and human placental lactogen were measured by the method described by Samaan et al. 17 The growth hormone was measured in 0.1 ml serum samples in duplicate. The addition of 1000 to 20,000 ng/ml of human placental lactogen in the assay tubes showed no cross-reaction with the growth hormone antibody used. The immunoreactive somatomedin C was measured in unextracted serum by the method of Furlanetto et al. 13 and in extracted serum by the method of the Immunonuclear Corporation, Sweetwater, Minnesota. In these somatomedin assays the addition of 1000 to 100,000 ng/ml of human placental lactogen to the incubating tubes did not interfere with the assay. In these immunoreactive methods, the intra-assay variation was ,,,;33 and the interassay variation was< 10%. The suppressible typical and nonsuppressible atypical insulin-
like activity was measured and transformed into units by the method of Samaan et al., 2 in quadruplicate aliquots of 0.25 ml serum. A reference serum sample (control) was included with each assay. The assay is sensitive to 4 µU of insulin. The interassay coefficient of variation on different days was < 10%. 2 The nonsuppressible insulin-like activity also was measured in five serum samples from two term normal pregnant women and three samples from diabetic pregnant women before and after acid alcohol extraction by the method of Daughaday et al. 18 to determine whether there was significant difference in the results obtained. Statistical analysis was done with use of the Student's t test for comparison and Pearson's r correlation coefficients. Data are presented as mean ± SD. Results
The maternal fasting blood glucose level in group C was 96 ± 19 mg/ 100 ml compared with 79 ± 15 mg/100 ml in group A (p < 0.02). The 2-hour blood glucose level after administration of 75 gm of oral glucose in group C was 198 ± 17 mg/100 ml compared with 142 ± 30 mg/100 ml in group A (p < 0.001). The birth weight of infants of diabetic mothers was 3962 ± 564 gm, which was significantly higher than that found in infants of nondiabetic pregnant mothers (3091 ± 519, p < 0.001 ). The average weight of term infants was 3091 ± 519 gm compared with 1650 ± 578 gm in premature infants (p < 0.001). The mean (±SD) maternal growth hormone levels were 1.95 ± 0.2 ng/ml, 2.3 ± 0.9 ng/ml, 2.1 ± 0.3, and 3.2 ± 2.1 ng/ml in groups A, B, C and D, respectively. Mean human placental lactogen levels were 6800 ± 2100 ng/ml, 5600 ± 1200 ng/ml, 8850 ± 1400 ng/ml (p < 0.001), and <20 ng/ml in groups A, B, C, and D, respectively. The values in groups B and C differed significantly (p < 0.001) from the level in group A. The human placental lactogen level in the umbilical cord of these infants was unmeasurable (<20 ng/ml). The somatomedin C in the extracted serum was 2.9 ± 0.9 IU/ml in group A, 2.4 ± 0.7 IU/ml in group B, and 2.8 ± 0.7 IU/ml in group C, all significantly higher (p < 0.001) than in group D (1.15 ± 0.37 IU/ml) (Table I). Similar relative results were obtained with use of unextracted serum, but the values were higher; these results were 8.5 ± 3.3 IU/ml, 7.4 ± 4.4 IU/ml, and 8.2 ± 3.1 IU/ml in groups A, B, and C, respectively, all being significantly higher than in group D (1.3 ± 0.3 IU/ml, p < 0.001). The mean growth hormone level in premature infants (group B, 23 ± 11 ng/ml) was significantly higher than that in group A infants (14 ± 7.1 ng/ml, p < 0.01). Both groups had higher values than normal adults (3.2 ± 2 .1 ng/ml, p < 0.001 ). In contrast, the mean cord blood somatomedin C levels were similar
Nonsuppressible insulin-like activity in mother and neonate
Volume 153 Number 4
459
Table I. Summary of the results of comparisons of the nonsuppressible insulin-like activity, somatomedin C level, growth hormone level, and birth weight for 17 normal pregnant women (38 to 41 weeks' gestation) and their fetuses at term (group A), 15 women (26 to 36 weeks' gestation) who delivered prematurely (group B), and 14 gestational diabetic women (37 to 41 weeks' gestation) and their fetuses (group C) compared with normal nonpregnant women (group D) Fetal
Maternal
Group
A B
c
D
p values A vs. B A vs. C Avs. D B vs. D C vs. D
N onsuppressible insulin-like activity (µ,u/ml)
291 209 224 164
± ± ± ±
89 69 77 12
<0.01 <0.05 <0.01 <0.05 <0.01
Growth hormone (nglml)
1.95 2.3 2.1 3.2
± ± ± ±
0.2 0.9 0.3 2.1
<0.2 <0.1 <0.05 <0.2 <0.01
Somatomedin C (JU/ml)
2.9 2.4 2.8 1.15
± ± ± ±
0.9 0.7 0.7 0.37
<0.5 <0.8 <0.001 <0.001 <0.05
m group A and B infants (0.9 ± 0.6 and 0.5 ± 0.3
IU/ml, p < 0.05). The mean somatomedin C level in infants of diabetic mothers was 1.3 ± 0. 7 IU/ml, a value higher than that found in the umbilical cord of group A normal neonates, but the difference did not reach significance (p < 0.1). The nonsuppressible insulin-like activity in premature infants was 130 ± 60 µU/ml, significantly less than the value in term infants (226 ± 76 µU/ml, p < 0.001). The results obtained from the acid alcohol extract of the serum of pregnant mothers with use of the paired t test were not significantly different from those obtained using unextracted serum (p < 0.9) (Table II). There was no statistically significant correlation between neonatal weight and maternal hormonal levels of nonsuppressible insulin-like activity, somatomedin C, human placental lactogen, or growth hormone. However, there was a positive correlation between maternal levels of nonsuppressible insulin-like activity and human placental lactogen in groups A and B (p < 0.05) but not in Group C.
Comment From these results as well as those of other studies12·15 it is apparent that maternal somatomedin C is significantly elevated during pregnancy. This increase occurs in spite of inhibition of growth hormone secretion during the third trimester. 19·21 Furthermore, Merimee et al. 22 showed that both insulin-like growth factors I and II were increased in the serum of a pregnant dwarf deficient in growth hormone. Daughaday and Kapadia 23 also showed that the somatomedin level, as measured by the costal cartilage assay, continues to rise in pregnant rats in spite of hypophysectomy. Our study shows that the nonsuppressible insulin-
N onsuppressible insulin-like activity (µ,Ulml)
Growth hormone (nglml)
S omatomedin C (JU/ml)
Birth weight
226 ± 76 130 ± 60 190 ± 58
14 ± 7.1 23 ± 11 15 ± 8.4
0.9 ± 0.6 0.5 ± 0.3 1.3 ± 0.7
3091 ± 519 1650 ± 578 3962 ± 564
<0.001 <0.3 <0.01 <0.05 <0.2
<0.01 <0.7 <0.001 <0.001 <0.001
<0.05 <0.1 <0.05 <0.001
<0.001 <0.001
(gm)
Table II. The nonsuppressible insulin-like activity in nonextracted and extracted serum from two term normal pregnant women (Nos. 1 and 2) and three term gestational diabetic women (Nos. 3, 4, and 5) Nonsuppressible insulin-like activity (µ,U/ml) Sample No. I
2 3 4
5
Mean
Nonextracted serum
Acid alcohol extract
415 381 191 191 164 268 ±SD 119
506 300 190 195 168 272 ± SD 141
p< 0.9.
like activity as measured by the fat pad assay also is significantly elevated during pregnancy. We also found that diabetic pregnant subjects have significantly lower nonsuppressible insulin-like activity than women during normal pregnancy (Table I). This result is at variance with the report of Herington et al., 15 who showed no difference. The difference between our results and those of Herington et al. 15 may be attributed to the technique used. Franklin et al 12 and Herington et al. 15 suggested that the high nonsuppressible insulin-like activity during pregnancy may be related directly to human placental lactogen levels. However, if this hypothesis is accepted, we would have seen higher nonsuppressible insulin-like activity levels in diabetic mothers in the presence of high human placental lactogen. We suggest that the increase of somatomedin C levels and nonsuppressible insulin-like activity during pregnancy may be related to the rise of insulin. During
460
Samaan et al.
pregnancy there is a progressive increase of insulin secretion in response to a glucose load or to arginine infusion. 20 · 21 These changes correspond well to the progressive rise of blood levels of human placental lactogen and complementary suppression of growth hormone. 21 Previously we showed that during the third trimester the peak immunoreactive insulin level was nearly tripled after glucose or arginine infusion compared to that found in the same patients 6 to 8 weeks after delivery!' We have also shown that women with gestational diabetes have lower immunoreactive insulin levels than normal pregnant subjects in spite of hyperglycemia. 21 Evidence was also presented that the rise of serum immunoreactive insulin levels in the mother during gestation is probably due to the presence of high human placental lactogen. 21 • 24 • 25 Based on these findings it appears that the significantly high levels of nonsuppressible insulin-like activity and somatomedin C in the mother during pregnancy may have resulted from the hyperinsulinism of pregnancy in the presence of inhibition of growth hormone secretion. The lower level of the nonsuppressible insulin-like activity seen in pregnant diabetic patients may be explained on the basis of the lower insulin secretion in spite of higher human placental Iactogen levels than seen in normal pregnant subjects. 21 The low level of nonsuppressible insulin-like activity in mothers who delivered prematurely also may be explained by the same phenomena, since the insulin level correlated with the period of gestation. 21 However, we cannot exclude the synergetic effect of human placental lactogen on insulin in the liver to release nonsuppressible insulin-like activity. There has been no general agreement regarding the hormones responsible for intrauterine fetal growth. Recently Gluckman et al. 16 measured the insulin-like growth factor I in the umbilical cord blood and showed results similar to ours for somatomedin C. They concluded that insulin-like growth factor I may play a role in fetal growth in spite of its significantly low levels. They also found that levels of insulin-like growth factor II measured by radioreceptor assay in the umbilical cord were similar to the adult values, but no correlation with gestational age or birth size was found. They also reviewed the literature, which showed a lack of general agreement on the level of insulin-like growth factor II in the umbilical cord. Our investigations showed that the nonsuppressible insulin-like activity as measured by the fat pad assay, which measures insulin-like growth factors I and II as well as unidentified insulin-like substances," was significantly higher in term infants than in premature infants (p < 0.001) (Table 1). Similarly, somatomedin C levels were higher in mature infants than in premature ones. However, growth hormone levels in the premature infants were higher than in term
October 15, 1985 Am J Obstet Gynecol
infants (p < 0.01), but both were significantly higher than in normal adults (p < 0.001), a finding which was · also reported by Cornblath et al. 26 The possible relationships of growth hormone, somatomedins, nonsuppressible insulin-like activity, and insulin to fetal growth have not been well delineated. Brinsmead and Liggins 27 reported that growth hormone is not required for fetal growth, since neither somatomedin levels nor growth are affected by hypophysectomy in rats or decapitation in the uterus. Costin et al. 28 reported that children with obesity of hyperphagia and catch-up growth after surgery for hypothalamic tumors had insulin levels that were normal or elevated and had normal somatomedin levels with good growth despite absent or low levels of growth hormone. From our studies we conclude that mothers with gestational diabetes have lower nonsuppressible insulinlike activity than normal pregnant mothers and that the nonsuppressible insulin-like activity level appears to correlate with the low level of immunoreactive insulin which is present in these patients in spite of higher human placental lactogen levels than those seen in normal pregnant women. The premature infants had lower nonsuppressible insulin-like activity levels than mature infants in spite of higher growth hormone levels. The nonsuppressible insulin-like activity and somatomedin C levels were not significantly elevated in children of diabetic mothers in spite of higher birth weights, indicating that the fetal overweight in this group of children may be related not only to nonsuppressible insulin-like activity and somatomedin C but also to the hyperinsulinemia known to be found in these infants, which is produced by the hyperglycemia. There was a paradoxical relationship between growth hormone and somatomedin C in the neonates. Both nonsuppressible insulin-like activity and somatomedin C levels are elevated in the pregnant mother in the presence of suppression of growth hormone. We wish to thank the National Pituitary Agency for supplying growth hormone and somatomedin C antibody for the radioimmunoassay, Dr. Mary A. Root of Eli Lilly for supplying insulin, and Professor R. E. Humbel of Switzerland for supplying somatomedin C. We thank Dr. Kuo-Pao Paul Yang for the measurement of somatomedin C levels in the extracted serum. We also thank Mr. Alonzo Guerra, Ms. Karen Smith, Ms. Socorro Castillo, and Ms. Dorothy Deinzer for their excellent technical assistance as well as Dr. Dennis Johnston from the Department of Biomathematics. REFERENCES l. Rinderknecht E, Humbel RE. Polypeptides with nonsuppressible insulin-like and cell-growth growth promoting activities in human serum: isolation, chemical character-
Volume 153 Number 4
2. 3.
4.
5. 6.
7. 8.
9.
10.
11.
12.
13. 14.
ization and some biological properties of forms I and II. Proc Natl Acad Sci USA 1976;73:2365. Samaan NA, Dempster WJ, Fraser R, Please NW, Stillman D. Further immunological studies on the form of circulating insulin. J Endocrinol 1962;24:263. Froesch ER, Burgi H, Ramseier EB, et al. Antibody-suppressible and nonsuppressible insulin-like activities in human serum and their physiologic significance: an insulin assay with adipose tissue of increased precision and specificity.] Clin Invest 1963;42:1816. Salmon WD Jr, Daughaday WH. Sulfation factor, a serum component mediating the action of growth hormone in stimulating incorporation of sulfate into cartilage. J Clin Invest l 956;35:733. Samaan NA, Stillman D, Fraser R. Abnormalities of serum insulin-like activity in liver disease. Lancet 1962;2:1287. Samaan NA, Dempster WJ, Fraser R, Stillman D. Changes in levels of "atypical" circulating insulin after infusing "typical" insulin through the liver. J Clin Endocrinol Metab 1963;26:1. Samaan NA, Fraser R, Dempster WJ. The "typical" and "atypical" forms of serum insulin. Diabetes 1963; 12:339. Daughaday WH, Phillips LS, Mueller MC. The effect of insulin and growth hormone on the release of somatomedin by the isolated rat liver. Endocrinology 1976;98:1214. Fryklund L, Skattner A, Hall K. Chemistry and biology of the somatomedins. In: Kasturp KW, Nielsen JK, eds. Proceedings of the eleventh annual FEBS meeting of the Federation of European Biochemical Societies. Vol. 28. New York: Pergamon Press, 1977:65. Underwood LE, Hintz RL, Voina SJ, Van WykJ.J. Human somatomedin, the growth hormone-dependent sulfation factor, is anti-lipolytic. J Clin Endocrinol Metab 1972;35:194. Zapf J, Rinderknecht E, Humbel RE, Froesch ER. Nonsuppressible insulin-like activity from human serum: recent accomplishments and their physiological implications. Metabolism 1978;27:1803. Franklin RC, Pepperell RJ, Rennie GC, Cameron DP. Acid-ethanol extractable nonsuppressible insulin-like activity (NSILA-S) during pregnancy and the puerperium, and in cord serum at term. J Clin Endocrinol Metab I 979;48:695. Furlanetto RW, Underwood LE, Van WykJJ, Handwerger S. Serum immunoreactive somatomedin-C is elevated late in pregnancy. J Clin Endocrinol Metab 1980;47:695. Sara VR, Hall K. Somatomedins and the fetus. Clin Obstet Gynecol I 980;23:765.
Nonsuppressible insulin-like activity in mother and neonate
461
15. Herington AC, Martin FIR, McBain JC. Nonsuppressible insulin-like activity during pregnancy in women with diabetes mellitus. Clin Endocrinol 1981;15:133. 16. Gluckman PD,Johnson-BarrettJJ, Butler JH, Edgar BW, Gunn TR: Studies of insulin-like growth factor-I and -II by specific radioligand assays in umbilical cord blood. Clin Endocrinol 1983; 19:405. 17. Samaan NA, Yen SCC, Friesen H, Pearson OH. Serum placental lactogen levels during pregnancy in trophoblastic disease. J Clin Endocrinol Metab 1966;26: 1303. 18. Daughaday WH, Mariz IK, Blethen SL. Inhibition of access of bound somatomedin to membrane receptor and immunobinding sites: a comparison of radioreceptor and radioimmunoassay of somatomedin in native and acid-ethanol-extracted serum. J Clin Endocrinol Metab 1980;5 I :781. 19. Yen SCC, Samaan NA, Pearson OH. Growth hormone levels in pregnancy. J Clin Endocrinol 1976;27: 1341. 20. Tyson JE, Rabinowitz D, Merimee TJ, Friesen H. Response of plasma insnlin and human growth hormone to arginine in pregnant and postpartum females. AM J 0BSTET GYNECOL 1969;103:313. 21. Samaan NA, McRoberts WA, Smith JP, Myers LG. Metabolic changes in women with trophoblastic disease and intrauterine fetal death compared with metabolic changes during normal pregnancy. J Clin Endocrinol Metab 1971;33:521. 22. Merimee TJ, Zapf J, Froesch ER. Insulin-like growth factor in pregnancy: studies in a growth hormone-deficient dwarf. J Clin Endocrinol Metab 1982;54: 110 I. 23. Daughaday WH, Kapadia M. Maintenance of serum somatomedin activity in hypophysectomized pregnant rats. Endocrinology 1978; 102: 1317. 24. Samaan NA, Yen SCC, Gonzalez D, Pearson OH. Metabolic effects of placental lactogen (HPL) in man. J Clin Endocrinol Metab l 968;28:485. 25. Beck P, Daughaday WH. Human placental lactogen: studies of its acute metabolic effects and disposition in normal man. J Clin Invest 1967;46:!03. 26. Cornblath M, Parker ML, Reisner SH, Forbes AE, Daughaday WH. Secretion and metabolism of growth hormone in premature and full term infants. J Clin Endocrinol Metab 1965 ;25 :209. 27. Brinsmead MW, Liggins GC. Serum somatomedin activity after hypophysectomy and during parturition in fetal lambs. Endocrinology 1979;105:297. 28. Costin G, Kogut M, Phillips LS, Daughaday WH. Craniopharyngioma: the role of insulin in promoting postoperative growth. J Clin Endocrinol Metab I 976;42:370.
Key words: Nonsuppressible insulin-like activity, somatomedins, mother, fetus
The mechanism responsible for fetal growth in utero is poorly understood. The isolation and the description of the sequence of the insulin-like growth factors I and II by Rinderknecht and Humbel in 1976 1 shed some light on this important subject. The isolation of these growth factors, also known as somatomedins, was preceded by the description of "typical" and "atypical" insulin-like activity by Samaan et al., 2 later called "suppressible" and "nonsuppressible" by Froesch et al., 3 and of a serum component (sulfation factor), mediating the growth hormone-induced incorporation of sulfate into cartilage by Salmon and Daughaday. 4 Numerous atFrom the Section of Endocrinology, Department of Medicine, The University of Texas M. D. Anderson Hospital and Tumor Institute at Houston, and the Department of Obstetrics and Gynecology, The University of Texas Medical School at Houston. This study was supported by Grant No. CA-35040 awarded by the National Cancer Institute and the Nancy Carmichael Gift Fund. Presented at the Thirty-second Annual Meeting of the Society for Gynecologic Investigation, Phoenix, Arizona, March 20-23, 1985. Reprint requests: Naguib A. Samaan, M.D., Ph.D., Section of Endocrinology, The University of Texas M. D. Anderson Hospital and Tumor Institute at Houston, 6723 Bertner Ave., Houston, TX 77030.
tempts have been made to identify these substances, their site of formation, and interrelationships. Samaan et al. 2 • 5 · 7 produced strong evidence that nonsuppressible insulin-like activity, or "atypical" insulin-like activity, was formed in the liver after regular insulin, "typical insulin-like activity," or insulin extracted from the pancreas was infused into the liver. Recently Daughaday et al. 8 also showed evidence that the liver plays a major part in formation of sulfation factor when growth hormone as well as insulin was infused into the liver, and they introduced the name somatomedin for the sulfation factor. Fryklund et al. 9 succeeded in purifying somatomedin A and Underwood et al. 10 purified and separated somatomedin C from adult human plasma. Rinderknecht and Humbel, 1 in addition to purifying insulin-like growth factors I and II, also determined their amino acid sequence, which was found to be similar to proinsulin. Zapf et al. 11 found that both insulin-like growth factors I and II fulfill all the criteria of the somatomedins. lnsulinlike growth factor I was reported to be structurally similar to somatomedin C, whereas insulin-like growth
457
458 Samaan et al.
October 15, 1985 Am J Obstet Gynecol
factor II represents a fraction of the nonsuppressible insulin-like activity. 11 Insulin-like growth factor II was found to be more potent than insulin-like growth factor I on adipose tissue but less potent in cartilage. 11 Somatomedin A and somatomedin C also were found to share the same biologic properties but to differ in their amino acid compositions. 11 Insulin-like growth factor I and II or the somatomedins have been measured in the sera of pregnant women and the umbilical cord by various techniques ~ ; however, the correlation between the somatomedins and nonsuppressible insulinlike activity as well as the hormones of pregnancy in mother and fetus need further clarification. This study was undertaken (1) to measure the maternal and fetal levels of somatomedin C and nonsuppressible insulin-like activity in relation to selected hormone levels at term and (2) to relate these biochemical findings with the clinical settings, in particular diabetes and premature birth. 12 16
Material and methods
Sera were collected by venipuncture at the time of delivery from 17 normal women at term (group A, gestation of 38 to 41 weeks), 15 normal women who delivered prematurely (group B, gestation of 26 to 36 weeks), and 14 women with gestational diabetes who did not receive insulin treatment but were treated with diet alone (group C, gestation of 37-41 weeks). Peripheral serum samples also were collected from 15 normal nonpregnant women matched for age (group D). Umbilical cord blood was obtained from groups A, B, and C at delivery immediately after the cord was clamped. The blood was collected in nonheparinized tubes from the placental side, which contained both arterial and venous blood. The serum was·separated and kept at - 20° C until the time of assay. This study was done with the approval of the Human Protection Committee, and informed consent was obtained from each mother. Immunoreactive growth hormone and human placental lactogen were measured by the method described by Samaan et al. 17 The growth hormone was measured in 0.1 ml serum samples in duplicate. The addition of 1000 to 20,000 ng/ml of human placental lactogen in the assay tubes showed no cross-reaction with the growth hormone antibody used. The immunoreactive somatomedin C was measured in unextracted serum by the method of Furlanetto et al. 13 and in extracted serum by the method of the Immunonuclear Corporation, Sweetwater, Minnesota. In these somatomedin assays the addition of 1000 to 100,000 ng/ml of human placental lactogen to the incubating tubes did not interfere with the assay. In these immunoreactive methods, the intra-assay variation was ,,,;33 and the interassay variation was< 10%. The suppressible typical and nonsuppressible atypical insulin-
like activity was measured and transformed into units by the method of Samaan et al., 2 in quadruplicate aliquots of 0.25 ml serum. A reference serum sample (control) was included with each assay. The assay is sensitive to 4 µU of insulin. The interassay coefficient of variation on different days was < 10%. 2 The nonsuppressible insulin-like activity also was measured in five serum samples from two term normal pregnant women and three samples from diabetic pregnant women before and after acid alcohol extraction by the method of Daughaday et al. 18 to determine whether there was significant difference in the results obtained. Statistical analysis was done with use of the Student's t test for comparison and Pearson's r correlation coefficients. Data are presented as mean ± SD. Results
The maternal fasting blood glucose level in group C was 96 ± 19 mg/ 100 ml compared with 79 ± 15 mg/100 ml in group A (p < 0.02). The 2-hour blood glucose level after administration of 75 gm of oral glucose in group C was 198 ± 17 mg/100 ml compared with 142 ± 30 mg/100 ml in group A (p < 0.001). The birth weight of infants of diabetic mothers was 3962 ± 564 gm, which was significantly higher than that found in infants of nondiabetic pregnant mothers (3091 ± 519, p < 0.001 ). The average weight of term infants was 3091 ± 519 gm compared with 1650 ± 578 gm in premature infants (p < 0.001). The mean (±SD) maternal growth hormone levels were 1.95 ± 0.2 ng/ml, 2.3 ± 0.9 ng/ml, 2.1 ± 0.3, and 3.2 ± 2.1 ng/ml in groups A, B, C and D, respectively. Mean human placental lactogen levels were 6800 ± 2100 ng/ml, 5600 ± 1200 ng/ml, 8850 ± 1400 ng/ml (p < 0.001), and <20 ng/ml in groups A, B, C, and D, respectively. The values in groups B and C differed significantly (p < 0.001) from the level in group A. The human placental lactogen level in the umbilical cord of these infants was unmeasurable (<20 ng/ml). The somatomedin C in the extracted serum was 2.9 ± 0.9 IU/ml in group A, 2.4 ± 0.7 IU/ml in group B, and 2.8 ± 0.7 IU/ml in group C, all significantly higher (p < 0.001) than in group D (1.15 ± 0.37 IU/ml) (Table I). Similar relative results were obtained with use of unextracted serum, but the values were higher; these results were 8.5 ± 3.3 IU/ml, 7.4 ± 4.4 IU/ml, and 8.2 ± 3.1 IU/ml in groups A, B, and C, respectively, all being significantly higher than in group D (1.3 ± 0.3 IU/ml, p < 0.001). The mean growth hormone level in premature infants (group B, 23 ± 11 ng/ml) was significantly higher than that in group A infants (14 ± 7.1 ng/ml, p < 0.01). Both groups had higher values than normal adults (3.2 ± 2 .1 ng/ml, p < 0.001 ). In contrast, the mean cord blood somatomedin C levels were similar
Nonsuppressible insulin-like activity in mother and neonate
Volume 153 Number 4
459
Table I. Summary of the results of comparisons of the nonsuppressible insulin-like activity, somatomedin C level, growth hormone level, and birth weight for 17 normal pregnant women (38 to 41 weeks' gestation) and their fetuses at term (group A), 15 women (26 to 36 weeks' gestation) who delivered prematurely (group B), and 14 gestational diabetic women (37 to 41 weeks' gestation) and their fetuses (group C) compared with normal nonpregnant women (group D) Fetal
Maternal
Group
A B
c
D
p values A vs. B A vs. C Avs. D B vs. D C vs. D
N onsuppressible insulin-like activity (µ,u/ml)
291 209 224 164
± ± ± ±
89 69 77 12
<0.01 <0.05 <0.01 <0.05 <0.01
Growth hormone (nglml)
1.95 2.3 2.1 3.2
± ± ± ±
0.2 0.9 0.3 2.1
<0.2 <0.1 <0.05 <0.2 <0.01
Somatomedin C (JU/ml)
2.9 2.4 2.8 1.15
± ± ± ±
0.9 0.7 0.7 0.37
<0.5 <0.8 <0.001 <0.001 <0.05
m group A and B infants (0.9 ± 0.6 and 0.5 ± 0.3
IU/ml, p < 0.05). The mean somatomedin C level in infants of diabetic mothers was 1.3 ± 0. 7 IU/ml, a value higher than that found in the umbilical cord of group A normal neonates, but the difference did not reach significance (p < 0.1). The nonsuppressible insulin-like activity in premature infants was 130 ± 60 µU/ml, significantly less than the value in term infants (226 ± 76 µU/ml, p < 0.001). The results obtained from the acid alcohol extract of the serum of pregnant mothers with use of the paired t test were not significantly different from those obtained using unextracted serum (p < 0.9) (Table II). There was no statistically significant correlation between neonatal weight and maternal hormonal levels of nonsuppressible insulin-like activity, somatomedin C, human placental lactogen, or growth hormone. However, there was a positive correlation between maternal levels of nonsuppressible insulin-like activity and human placental lactogen in groups A and B (p < 0.05) but not in Group C.
Comment From these results as well as those of other studies12·15 it is apparent that maternal somatomedin C is significantly elevated during pregnancy. This increase occurs in spite of inhibition of growth hormone secretion during the third trimester. 19·21 Furthermore, Merimee et al. 22 showed that both insulin-like growth factors I and II were increased in the serum of a pregnant dwarf deficient in growth hormone. Daughaday and Kapadia 23 also showed that the somatomedin level, as measured by the costal cartilage assay, continues to rise in pregnant rats in spite of hypophysectomy. Our study shows that the nonsuppressible insulin-
N onsuppressible insulin-like activity (µ,Ulml)
Growth hormone (nglml)
S omatomedin C (JU/ml)
Birth weight
226 ± 76 130 ± 60 190 ± 58
14 ± 7.1 23 ± 11 15 ± 8.4
0.9 ± 0.6 0.5 ± 0.3 1.3 ± 0.7
3091 ± 519 1650 ± 578 3962 ± 564
<0.001 <0.3 <0.01 <0.05 <0.2
<0.01 <0.7 <0.001 <0.001 <0.001
<0.05 <0.1 <0.05 <0.001
<0.001 <0.001
(gm)
Table II. The nonsuppressible insulin-like activity in nonextracted and extracted serum from two term normal pregnant women (Nos. 1 and 2) and three term gestational diabetic women (Nos. 3, 4, and 5) Nonsuppressible insulin-like activity (µ,U/ml) Sample No. I
2 3 4
5
Mean
Nonextracted serum
Acid alcohol extract
415 381 191 191 164 268 ±SD 119
506 300 190 195 168 272 ± SD 141
p< 0.9.
like activity as measured by the fat pad assay also is significantly elevated during pregnancy. We also found that diabetic pregnant subjects have significantly lower nonsuppressible insulin-like activity than women during normal pregnancy (Table I). This result is at variance with the report of Herington et al., 15 who showed no difference. The difference between our results and those of Herington et al. 15 may be attributed to the technique used. Franklin et al 12 and Herington et al. 15 suggested that the high nonsuppressible insulin-like activity during pregnancy may be related directly to human placental lactogen levels. However, if this hypothesis is accepted, we would have seen higher nonsuppressible insulin-like activity levels in diabetic mothers in the presence of high human placental lactogen. We suggest that the increase of somatomedin C levels and nonsuppressible insulin-like activity during pregnancy may be related to the rise of insulin. During
460
Samaan et al.
pregnancy there is a progressive increase of insulin secretion in response to a glucose load or to arginine infusion. 20 · 21 These changes correspond well to the progressive rise of blood levels of human placental lactogen and complementary suppression of growth hormone. 21 Previously we showed that during the third trimester the peak immunoreactive insulin level was nearly tripled after glucose or arginine infusion compared to that found in the same patients 6 to 8 weeks after delivery!' We have also shown that women with gestational diabetes have lower immunoreactive insulin levels than normal pregnant subjects in spite of hyperglycemia. 21 Evidence was also presented that the rise of serum immunoreactive insulin levels in the mother during gestation is probably due to the presence of high human placental lactogen. 21 • 24 • 25 Based on these findings it appears that the significantly high levels of nonsuppressible insulin-like activity and somatomedin C in the mother during pregnancy may have resulted from the hyperinsulinism of pregnancy in the presence of inhibition of growth hormone secretion. The lower level of the nonsuppressible insulin-like activity seen in pregnant diabetic patients may be explained on the basis of the lower insulin secretion in spite of higher human placental Iactogen levels than seen in normal pregnant subjects. 21 The low level of nonsuppressible insulin-like activity in mothers who delivered prematurely also may be explained by the same phenomena, since the insulin level correlated with the period of gestation. 21 However, we cannot exclude the synergetic effect of human placental lactogen on insulin in the liver to release nonsuppressible insulin-like activity. There has been no general agreement regarding the hormones responsible for intrauterine fetal growth. Recently Gluckman et al. 16 measured the insulin-like growth factor I in the umbilical cord blood and showed results similar to ours for somatomedin C. They concluded that insulin-like growth factor I may play a role in fetal growth in spite of its significantly low levels. They also found that levels of insulin-like growth factor II measured by radioreceptor assay in the umbilical cord were similar to the adult values, but no correlation with gestational age or birth size was found. They also reviewed the literature, which showed a lack of general agreement on the level of insulin-like growth factor II in the umbilical cord. Our investigations showed that the nonsuppressible insulin-like activity as measured by the fat pad assay, which measures insulin-like growth factors I and II as well as unidentified insulin-like substances," was significantly higher in term infants than in premature infants (p < 0.001) (Table 1). Similarly, somatomedin C levels were higher in mature infants than in premature ones. However, growth hormone levels in the premature infants were higher than in term
October 15, 1985 Am J Obstet Gynecol
infants (p < 0.01), but both were significantly higher than in normal adults (p < 0.001), a finding which was · also reported by Cornblath et al. 26 The possible relationships of growth hormone, somatomedins, nonsuppressible insulin-like activity, and insulin to fetal growth have not been well delineated. Brinsmead and Liggins 27 reported that growth hormone is not required for fetal growth, since neither somatomedin levels nor growth are affected by hypophysectomy in rats or decapitation in the uterus. Costin et al. 28 reported that children with obesity of hyperphagia and catch-up growth after surgery for hypothalamic tumors had insulin levels that were normal or elevated and had normal somatomedin levels with good growth despite absent or low levels of growth hormone. From our studies we conclude that mothers with gestational diabetes have lower nonsuppressible insulinlike activity than normal pregnant mothers and that the nonsuppressible insulin-like activity level appears to correlate with the low level of immunoreactive insulin which is present in these patients in spite of higher human placental lactogen levels than those seen in normal pregnant women. The premature infants had lower nonsuppressible insulin-like activity levels than mature infants in spite of higher growth hormone levels. The nonsuppressible insulin-like activity and somatomedin C levels were not significantly elevated in children of diabetic mothers in spite of higher birth weights, indicating that the fetal overweight in this group of children may be related not only to nonsuppressible insulin-like activity and somatomedin C but also to the hyperinsulinemia known to be found in these infants, which is produced by the hyperglycemia. There was a paradoxical relationship between growth hormone and somatomedin C in the neonates. Both nonsuppressible insulin-like activity and somatomedin C levels are elevated in the pregnant mother in the presence of suppression of growth hormone. We wish to thank the National Pituitary Agency for supplying growth hormone and somatomedin C antibody for the radioimmunoassay, Dr. Mary A. Root of Eli Lilly for supplying insulin, and Professor R. E. Humbel of Switzerland for supplying somatomedin C. We thank Dr. Kuo-Pao Paul Yang for the measurement of somatomedin C levels in the extracted serum. We also thank Mr. Alonzo Guerra, Ms. Karen Smith, Ms. Socorro Castillo, and Ms. Dorothy Deinzer for their excellent technical assistance as well as Dr. Dennis Johnston from the Department of Biomathematics. REFERENCES l. Rinderknecht E, Humbel RE. Polypeptides with nonsuppressible insulin-like and cell-growth growth promoting activities in human serum: isolation, chemical character-
Volume 153 Number 4
2. 3.
4.
5. 6.
7. 8.
9.
10.
11.
12.
13. 14.
ization and some biological properties of forms I and II. Proc Natl Acad Sci USA 1976;73:2365. Samaan NA, Dempster WJ, Fraser R, Please NW, Stillman D. Further immunological studies on the form of circulating insulin. J Endocrinol 1962;24:263. Froesch ER, Burgi H, Ramseier EB, et al. Antibody-suppressible and nonsuppressible insulin-like activities in human serum and their physiologic significance: an insulin assay with adipose tissue of increased precision and specificity.] Clin Invest 1963;42:1816. Salmon WD Jr, Daughaday WH. Sulfation factor, a serum component mediating the action of growth hormone in stimulating incorporation of sulfate into cartilage. J Clin Invest l 956;35:733. Samaan NA, Stillman D, Fraser R. Abnormalities of serum insulin-like activity in liver disease. Lancet 1962;2:1287. Samaan NA, Dempster WJ, Fraser R, Stillman D. Changes in levels of "atypical" circulating insulin after infusing "typical" insulin through the liver. J Clin Endocrinol Metab 1963;26:1. Samaan NA, Fraser R, Dempster WJ. The "typical" and "atypical" forms of serum insulin. Diabetes 1963; 12:339. Daughaday WH, Phillips LS, Mueller MC. The effect of insulin and growth hormone on the release of somatomedin by the isolated rat liver. Endocrinology 1976;98:1214. Fryklund L, Skattner A, Hall K. Chemistry and biology of the somatomedins. In: Kasturp KW, Nielsen JK, eds. Proceedings of the eleventh annual FEBS meeting of the Federation of European Biochemical Societies. Vol. 28. New York: Pergamon Press, 1977:65. Underwood LE, Hintz RL, Voina SJ, Van WykJ.J. Human somatomedin, the growth hormone-dependent sulfation factor, is anti-lipolytic. J Clin Endocrinol Metab 1972;35:194. Zapf J, Rinderknecht E, Humbel RE, Froesch ER. Nonsuppressible insulin-like activity from human serum: recent accomplishments and their physiological implications. Metabolism 1978;27:1803. Franklin RC, Pepperell RJ, Rennie GC, Cameron DP. Acid-ethanol extractable nonsuppressible insulin-like activity (NSILA-S) during pregnancy and the puerperium, and in cord serum at term. J Clin Endocrinol Metab I 979;48:695. Furlanetto RW, Underwood LE, Van WykJJ, Handwerger S. Serum immunoreactive somatomedin-C is elevated late in pregnancy. J Clin Endocrinol Metab 1980;47:695. Sara VR, Hall K. Somatomedins and the fetus. Clin Obstet Gynecol I 980;23:765.
Nonsuppressible insulin-like activity in mother and neonate
461
15. Herington AC, Martin FIR, McBain JC. Nonsuppressible insulin-like activity during pregnancy in women with diabetes mellitus. Clin Endocrinol 1981;15:133. 16. Gluckman PD,Johnson-BarrettJJ, Butler JH, Edgar BW, Gunn TR: Studies of insulin-like growth factor-I and -II by specific radioligand assays in umbilical cord blood. Clin Endocrinol 1983; 19:405. 17. Samaan NA, Yen SCC, Friesen H, Pearson OH. Serum placental lactogen levels during pregnancy in trophoblastic disease. J Clin Endocrinol Metab 1966;26: 1303. 18. Daughaday WH, Mariz IK, Blethen SL. Inhibition of access of bound somatomedin to membrane receptor and immunobinding sites: a comparison of radioreceptor and radioimmunoassay of somatomedin in native and acid-ethanol-extracted serum. J Clin Endocrinol Metab 1980;5 I :781. 19. Yen SCC, Samaan NA, Pearson OH. Growth hormone levels in pregnancy. J Clin Endocrinol 1976;27: 1341. 20. Tyson JE, Rabinowitz D, Merimee TJ, Friesen H. Response of plasma insnlin and human growth hormone to arginine in pregnant and postpartum females. AM J 0BSTET GYNECOL 1969;103:313. 21. Samaan NA, McRoberts WA, Smith JP, Myers LG. Metabolic changes in women with trophoblastic disease and intrauterine fetal death compared with metabolic changes during normal pregnancy. J Clin Endocrinol Metab 1971;33:521. 22. Merimee TJ, Zapf J, Froesch ER. Insulin-like growth factor in pregnancy: studies in a growth hormone-deficient dwarf. J Clin Endocrinol Metab 1982;54: 110 I. 23. Daughaday WH, Kapadia M. Maintenance of serum somatomedin activity in hypophysectomized pregnant rats. Endocrinology 1978; 102: 1317. 24. Samaan NA, Yen SCC, Gonzalez D, Pearson OH. Metabolic effects of placental lactogen (HPL) in man. J Clin Endocrinol Metab l 968;28:485. 25. Beck P, Daughaday WH. Human placental lactogen: studies of its acute metabolic effects and disposition in normal man. J Clin Invest 1967;46:!03. 26. Cornblath M, Parker ML, Reisner SH, Forbes AE, Daughaday WH. Secretion and metabolism of growth hormone in premature and full term infants. J Clin Endocrinol Metab 1965 ;25 :209. 27. Brinsmead MW, Liggins GC. Serum somatomedin activity after hypophysectomy and during parturition in fetal lambs. Endocrinology 1979;105:297. 28. Costin G, Kogut M, Phillips LS, Daughaday WH. Craniopharyngioma: the role of insulin in promoting postoperative growth. J Clin Endocrinol Metab I 976;42:370.