TNF-␣ Is a Predictor of Insulin Resistance in Human Pregnancy John P. Kirwan,1,2 Sylvie Hauguel-De Mouzon,3 Jacques Lepercq,3 Jean-Claude Challier,4 Larraine Huston-Presley,1 Jacob E. Friedman,5 Satish C. Kalhan,6 and Patrick M. Catalano1,6 Historically, insulin resistance during pregnancy has most significant independent predictor of insulin sensi- ؊0.67, P < 0.0001), even after adjustment for؍ tivity (r ؍ -been ascribed to increased production of placental hor mones and cortisol. The purpose of this study was to fat mass by covariance (r 0.46, P < 0.01). These test this hypothesis by correlating the longitudinal observations challenge the view that the classical repro- changes in insulin sensitivity during pregnancy with ductive hormones are the primary mediators of change in insulin sensitivity during gestation and provide the changes in placental hormones, cortisol, leptin, and ␣ tumor necrosis factor (TNF)-␣. Insulin resistance was basis for including TNF- in a new paradigm to explain assessed in 15 women (5 with gestational diabetes insulin resistance in pregnancy. Diabetes 51:2207–2213, mellitus [GDM] and 10 with normal glucose tolerance) 2002 using the euglycemic-hyperinsulinemic clamp proce- dure, before pregnancy (pregravid) and during early (12–14 weeks) and late (34–36 weeks) gestation. Body composition, plasma TNF-␣, leptin, cortisol, and repro- regnancy is a period marked by profound ductive hormones (human chorionic gonadotropin, es- changes in a woman’s hormonal status and me- tradiol, progesterone, human placental lactogen, and tabolism. The ability to regulate nutrient balance prolactin) were measured in conjunction with the during this period is critical to the health of the ␣ P clamps. Placental TNF- was measured in vitro using mother and the growing fetus. Insulin is one of the key dually perfused human placental cotyledon from five regulators of metabolism, and significant changes in insu- additional subjects. Compared with pregravid, insulin lin sensitivity and its ability to control glucose, fat, and resistance was evident during late pregnancy in all protein during pregnancy have been well documented women (12.4 ؎ 1.2 vs. 8.1 ؎ 0.8 10؊2 mg ⅐ kg؊1 fat-free ؊ ؊ ؊ (1–4). Previous reports have also shown that maternal mass ⅐ min 1 ⅐ ␮U 1 ⅐ ml 1). TNF-␣, leptin, cortisol, all reproductive hormones, and fat mass were increased in insulin resistance plays an important role in the regulation late pregnancy (P < 0.001). In vitro, most of the placen- of maternal energy metabolism, fat accretion, and fetal tal TNF-␣ (94%) was released into the maternal circu- growth (5,6). In gestational diabetes mellitus (GDM), lation; 6% was released to the fetal side. During late greater insulin resistance may lead to abnormal blood pregnancy, TNF-␣ was inversely correlated with insulin glucose and fetal macrosomia, may increase the likelihood -؊0.69, P < 0.006). Furthermore, among of obstetric complications, and in some cases, may in؍ sensitivity (r all of the hormonal changes measured in this study, the crease the risk of stillbirth. The cellular mediators of change in TNF-␣ from pregravid to late pregnancy was insulin resistance in late pregnancy have long been as- the only significant predictor of the change in insulin ؊ cribed to alterations in cortisol and placental-derived؍ sensitivity (r 0.60, P < 0.02). The placental repro- hormones including human placental lactogen (HPL), pro- ductive hormones and cortisol did not correlate with insulin sensitivity in late pregnancy. Multivariate step- gesterone, and estrogen (7–9). To our knowledge, how- wise regression analysis revealed that TNF-␣ was the ever, the changes in insulin sensitivity during gestation have not yet been correlated with hormonal changes using a prospective longitudinal design. From the 1Department of Reproductive Biology, Case Western Reserve Recently, investigators have focused on several new University School of Medicine at MetroHealth Medical Center, Cleveland, Ohio; the 2Department of Nutrition, Case Western Reserve University School potential mediators of insulin resistance including the of Medicine at MetroHealth Medical Center, Cleveland, Ohio; the 3De´parte- cytokine tumor necrosis factor (TNF)-␣, the anti-obesity ment d’Endocrinologie, Institut Cochin de Ge´ne´tique Mole´culaire, Paris, France; the 4Service de Gyne´cologie-Obste´trique, Hoˆ pital Cochin-Saint Vin- hormone leptin, and adipose-derived resistin, adiponectin, cent-de-Paul, Paris, France; the 5Departments of Pediatrics, Biochemistry, and and free fatty acids (10,11). Among these candidates, Molecular Genetics, University of Colorado Health Sciences Center, Denver, TNF-␣ and leptin are known to be produced in the Colorado; and the 6Schwartz Center for Metabolism and Nutrition, Case Western Reserve University School of Medicine at MetroHealth Medical placenta and therefore could play a central role in insulin Center, Cleveland, Ohio. resistance during pregnancy (12–14). Although increased Address correspondence and reprint requests to John P. Kirwan, Depart- circulating TNF-␣ levels have been associated with insulin ments of Reproductive Biology and Nutrition, Case Western Reserve University School of Medicine at MetroHealth Medical Center, Bell Greve resistance in obesity, aging, sepsis, muscle damage, and Bldg., Rm. G-232E, 2500 MetroHealth Dr., Cleveland, OH 44109-1998. preeclamptic pregnancy, reports of change in levels during E-mail: [email protected]. Received for publication 1 March 2002 and accepted in revised form 3 April normal pregnancy and GDM are equivocal (15–22). Fur- 2002. thermore, because TNF-␣ is synthesized and secreted FFM, fat-free mass; GDM, gestational diabetes mellitus; hCG, human from the placenta and adipose tissue, the origin of circu- chorionic gonadotropin; HPL, human placental lactogen; IRS, insulin receptor substrate; NGT, normal glucose tolerance; RIA, radioimmunoassay; TNF, lating levels during pregnancy remains largely unknown. tumor necrosis factor. Leptin was first identified as a product of the ob gene in DIABETES, VOL. 51, JULY 2002 2207 TNF-␣ AND INSULIN RESISTANCE DURING PREGNANCY TABLE 1 recirculation period were recorded, and the media were stored at Ϫ20°C for Subject characteristics and glucose tolerance pregravid and later analysis. To estimate TNF-␣ release into the fetal (Qf ϭ Cf ϫ Vf) and during late pregnancy maternal (Qm ϭ Cm ϫ Vm) circulation, TNF-␣ concentrations were measured in fetal (Cf) and maternal (Cm) media. These values were divided by the Lean NGT Obese NGT GDM duration of the perfusion (t ϭ 120 min) to give the rates of TNF-␣ release toward the fetal and maternal side of the placental circulation. TNF-␣ was also Age (years) 33 Ϯ 230Ϯ 129Ϯ 2 assessed on triplicate tissue samples in the placenta, before (Qti), and at the Parity 1.2 Ϯ 0.4 0.8 Ϯ 0.2 1.0 Ϯ 0.3 end of (Qte) the perfusion to evaluate changes in placental TNF-␣ (Q ϭ Qte Ϫ Weight (kg) 53.5 Ϯ 1.4 76.8 Ϯ 7.4* 83.3 Ϯ 9.4* Qti). The rate of placental TNF-␣ synthesis was estimated as PLS ϭ Qf ϩ Qm BMI (kg/m2) 19.8 Ϯ 1.0 27.3 Ϯ 2.4* 30.8 Ϯ 2.8* ϩ Q divided by the duration of the perfusion for one cotyledon, and as TPS ϭ Body fat (%) 25.0 Ϯ 1.6 36.6 Ϯ 3.2* 40.1 Ϯ 3.2* PLS ϫ P/l, where P/l is the placenta/lobule weight ratio, for the whole 2-h Glucose, placenta. ␣ pregravid 102.6 Ϯ 17.8 106.6 Ϯ 2.8 141.0 Ϯ 6.5 Analytical procedures. All samples for TNF- , leptin, cortisol, and the ␣ 2-h Glucose, during reproductive hormones were run in duplicate in a single assay. Plasma TNF- Ϯ Ϯ Ϯ concentrations were measured by enzyme-linked immunosorbent assay late pregnancy 125.0 8.3 145.6 8.1† 173.4 4.9*† (Quantikine HS; R&D Systems, Minneapolis, MN). The intra-assay coefficient Data are means Ϯ SE; n ϭ 5 per group. *Significantly different from of variation was 14%, the minimum detectable limit of the assay was 0.18 lean NGT group, P Ͻ 0.02. †Significantly different from pregravid, pg/ml, and the lowest standard was 0.5 pg/ml. Plasma glucose concentrations P Ͻ 0.03. were measured by the glucose oxidase method (Yellow Springs Instruments, Yellow Springs, OH). Blood samples for insulin measurements were centri- fuged at 4°C, and the plasma was stored at –70°C for subsequent analysis by adipose tissue and is thought to regulate energy balance a double-antibody radioimmunoassay (RIA) as previously described (27). through central hypothalamic pathways, but it may also Plasma cortisol, estradiol, HPL, and progesterone were determined by RIA play a role in insulin resistance by promoting lipid oxida- (Diagnostic Products, Los Angeles, CA). Plasma leptin samples were also measured by RIA (Linco Research, St. Charles, MO). Plasma hCG was tion and inhibiting lipid synthesis (23–25). There are determined by immunoradiometric assay (Diagnostic Products). Plasma pro- several reports of increased circulating leptin during preg- lactin was measured by RIA (Nichols Institute Diagnostics, San Juan Capist- nancy, due in part to placental secretion (13,14,26). rano, CA) 2 The primary purpose of this prospective study was to The [6,6- H2]glucose in the plasma samples was isolated by ion-exchange relate the longitudinal changes in concentrations of repro- chromatography. A penta-acetate derivative of glucose was prepared accord- ing to Tserng and Kalhan (33). Plasma enrichment was determined using a gas ductive hormones (HPL, progesterone, human chorionic chromatograph mass spectrometer (model 5985B; Hewlett-Packard, Palo Alto, gonadotropin [hCG], prolactin, and estradiol), TNF-␣, lep- CA). tin, and cortisol with the corresponding changes in mater- Calculations and statistical analysis. The insulin sensitivity index from the nal insulin sensitivity as measured by the euglycemic- clamp procedure was estimated as the glucose infusion rate plus endogenous hyperinsulinemic clamp. To further elucidate the role of glucose output divided by the mean insulin concentration during the clamp and is expressed as 10Ϫ2 milligrams per kilogram fat-free mass (FFM) per TNF-␣ as a mediator of maternal insulin resistance, we Ϫ Ϫ Ϫ Ϫ minute per microunit per milliliter (10 2 mg ⅐ kg 1 FFM ⅐ min 1/␮U ⅐ ml 1).