Escitalopram Ameliorates Hypercortisolemia and Insulin Resistance in Low Birth Weight Men With Limbic Brain Alterations

Escitalopram Ameliorates Hypercortisolemia and Insulin Resistance in Low Birth Weight Men With... Abstract Context Low birth weight (LBW; <2500 g) is linked to the development of insulin resistance and limbic-hypothalamic-pituitary-adrenal (LHPA) axis hyperactivity. Objective Our first aim was to study insulin action, LHPA axis function, and limbic brain structures in young, healthy LBW men vs normal birthweight (NBW) controls (part 1). Our second aim was to investigate the effects of escitalopram vs placebo in LBW men in the LHPA axis and insulin sensitivity (part 2). Design Setting, Participants, and Intervention The maximal (Rdmax) and submaximal (Rdsubmax) rates of insulin-stimulated glucose turnover, LHPA axis, and brain morphology were examined in 40 LBW men and 20 matched NBW men using two-stage hyperinsulinemic euglycemic clamp, 24-hour hormone plasma profiles, and magnetic resonance imaging. Subsequently, all LBW subjects underwent randomized and double-blind treatment with escitalopram 20 mg/d or placebo for 3 months followed by a complete reexamination. Main Outcome Measures (Part 2) Changes in Rdmax/Rdsubmax and plasma-free cortisol 24-hour area under the curve. Results In LBW vs NBW, Rdsubmax and Rdmax were ∼16% (P = 0.01) and ∼12% (P = 0.01) lower, respectively, and 24-hour free cortisol levels were ∼20% higher (P = 0.02), primarily driven by a ∼99% increase at 05:00 am (P < 0.001). Furthermore, these changes were related to structural alterations within left thalamus and ventromedial prefrontal cortex. However, in LBW men, exposure to escitalopram normalized the free cortisol levels and improved the Rdsubmax by ∼24% (P = 0.04) compared with placebo. Conclusions LBW vs NBW displayed alterations in key brain structures modulating the LHPA axis, elevated free cortisol levels, and insulin resistance. Escitalopram administration ameliorated these defects, suggesting a potential for LHPA axis modulation compounds to improve insulin action in LBW subjects. Low birthweight (LBW; <2500 g) is a crude marker of a compromised intrauterine environment and has consistently been associated with insulin resistance and an increased risk of type 2 diabetes (T2D) in adult life (1). However, the exact mechanisms responsible for these changes remain unknown. The “thrifty phenotype hypothesis” (2) suggests that fetal stress or insufficient fetal nutrition induce physiological adaptations to rescue fetus survival and to prepare its organism for a hostile postnatal environment. Such adaptations, however, might simultaneously increase the risk of various metabolic and neurocognitive disturbances in adult life (2, 3). Findings from both humans and animals have indicated that LBW can cause disturbances in the limbic–hypothalamic–pituitary–adrenal (LHPA) axis regarding both diurnal and stress-related homeostasis, resulting in overall greater circulating levels of cortisol (4–6). Overexposure to corticosteroids is known to impair tissue insulin sensitivity (7); hence, LHPA axis hyperactivity might be an important factor in the development of insulin resistance and T2D in adult subjects born with LBW (5). Magnetic resonance imaging (MRI) studies of children and adults with LBW have revealed changes in limbic structures such as the hippocampus, thalamus, and prefrontal cortex (8–10), which are key limbic organs known to be involved in LHPA axis regulation (11, 12). These studies, however, did not assess LHPA axis function, and it remains unknown whether structural limbic changes are directly related to neuroendocrine disturbances within the LHPA axis. LHPA axis hyperactivity has been related to major depressive disorder, and in these patients, treatment with a selective serotonin reuptake inhibitor (SSRI) has been found to downregulate LHPA axis tonus (13, 14) and improve insulin sensitivity (15, 16). Furthermore, in LBW rats characterized by LHPA axis hyperactivity and whole body insulin resistance, we have previously shown that 4 to 5 weeks of escitalopram treatment led to complete normalization of LHPA axis regulation and whole body insulin action (17, 18). Similar studies, however, have not yet been conducted in humans with LBW. The primary aim of the present investigation was to study the limbic brain structures, LHPA axis function, and insulin sensitivity in young men with LBW vs young men with normal birthweight (NBW; part 1, the LBW phenotype study). Our secondary aim was to assess the effects of escitalopram or placebo on the LHPA axis and insulin sensitivity in LBW subjects (part 2, the LBW trial). For part 1, we recruited 40 healthy LBW men and 20 NBW men matched by body mass index (BMI), age, and physical activity. After completion of part 1, all 40 LBW subjects were enrolled in a randomized double-blinded trial with 3 months of treatment with escitalopram (20 mg/d) or placebo, followed by a complete reexamination (part 2). In both part 1 and part 2, we examined LHPA axis regulation (24-hour plasma profiles), limbic brain morphology (MRI), and insulin sensitivity (hyperinsulinemic euglycemic clamp). Materials and Methods Participants We identified and recruited 40 male LBW subjects (<2500 g) born at term through the Danish Birth Register. We prospectively recruited and matched 20 male subjects with NBW (3500 to 4500 g) at term, matched by age, BMI, and physical activity, using local and national advertising. All 60 subjects were young, healthy, and nonsmokers. They were also lean (BMI <25 kg/m2), received no medication, had no history of depression, and no first-degree relatives with T2D. All participants were white, and the parents of the subjects were of normal height (father ≥170 cm, mother ≥160 cm). Study design The present investigation consisted of two studies: the LBW phenotype study (part 1) and the LBW trial (part 2). The complete experiments (parts 1 and 2) were performed from March to September to minimize any seasonal variations. In part 1, the LBW and NBW subjects were examined at two different visits, 1 week apart. Both visits lasted for 2 days and included 2 overnight stays. All visits were initiated by an overnight stay at the clinic before the experiments to acclimatize the participants to the hospital environment. All participants received standardized meals at every visit (at 8:30 am, 12:30 pm, and 6:30 pm) and were asked to abstain from alcohol use and strenuous physical activity for 4 days before each visit. In part 2, after completion of part 1, all LBW subjects were randomized to receive double-blind treatment with placebo or escitalopram for 3 months. The dosage was 10 mg/d for week 1 and escalated to a final dose of 20 mg/d at week 2. After 3 months of intervention (placebo 101.2 ± 1.8 days, escitalopram 102.1 ± 1.9 days), all LBW participants were reexamined. All reexamination procedures and visits were performed identically to those conducted at baseline in part 1. Telephone consultations were performed after 1, 2, 4, 8, and 12 weeks to monitor for adverse events and determine drug adherence. The randomization code (generated and kept by the hospital pharmacy) was not unblinded until after the last Good Clinical Practice (GCP) audit and after the last visit of the last patient. The primary endpoints were changes in maximal insulin-stimulated glucose turnover (Rdmax), submaximal insulin-stimulated glucose turnover (Rdsubmax), and plasma-free cortisol 24-hour area under the curve (AUC). Study approval All participants submitted written informed consent before study inclusion. The study was executed in accordance with the GCP guidelines and the Declaration of Helsinki, monitored by the local GCP unit at Aarhus University Hospital, and approved by the regional ethical committees, the Danish Health and Medicines Authority, and the Danish Data Protection Agency (approval no. EudraCT 2008-004521-42; ClinicalTrials.gov identifier, NCT00971815). LHPA axis function The diurnal LHPA axis activity was evaluated using 24-hour plasma profiles of adrenocorticotropic hormone (ACTH) and cortisol, with blood samples taken using an intravenous catheter at 8:00 am, 11:00 am, 2:00 pm, 5:00 pm, 8:00 pm, 10:00 pm, 11:00 pm, 12:00 am, 2:00 am, and 5:00 am. Corticosteroid-binding globulin (CBG) was detected to calculate the plasma concentrations of free cortisol according to Coolens equation, assuming constant CBG levels (19). The total 24-hour AUC was calculated using the trapezoid method (20). Cortisol was measured by the Clinical Biochemical Department, Aarhus University Hospital, using an electrochemiluminescence immunoassay [Cobas 6000 Analyzer; Roche, Hvidovre, Denmark; coefficient of variation (intermediate precision) 3.3%]. CBG and ACTH were analyzed using commercial enzyme-linked immunosorbent assay kits EIA-3647 (DRG Instruments GmbH, Marburg, Germany) and RD19223400R (Biovendor, Brno, Czech Republic). Hyperinsulinemic euglycemic clamp Clamps were initiated by primed (20 μCi) continuous infusion (0.20 μCi/min) of D-[3-3H] glucose for 2 hours to assess the basal glucose turnover. Next, glucose turnover was assessed at both submaximal (insulin infusion rate 0.3 mU/kg/min; Actrapid; Novo Nordisk A/S, Bagsvaerd, Denmark) and maximal (1.0 mU insulin/kg/min) insulin stimulation, with a duration of 2 hours for each of the two stages. Plasma glucose was maintained during insulin stimulation at 5.0 mM using variable 20% glucose infusion. At submaximal insulin stimulation, D-[3-3H] glucose tracer was added to the infused glucose (∼100 μCi/500 mL 20% glucose) to estimate the rate of endogenous glucose production (EGP) and rate of glucose disappearance (Rd) using the Steele correction for nonsteady state (21). The basal hepatic insulin sensitivity index (HSI) was calculated as described by Miyazaki et al. (22). Plasma glucose levels were analyzed in real time using a YSI 2300 analyzer. Steady-state plasma insulin levels were assessed after clamp using a time-resolved fluoroimmunoassay (Perkin Elmer, Skovlunde, Denmark). Serum-specific activity of D-[3-3H] glucose was measured using a beta counter (TRi-Carb 2910 TR; Perkin Elmer). Body composition and MRI lipid spectroscopy The body weights and waist/hip ratios were measured by trained laboratory staff. The whole body fat content and distribution were evaluated using the same dual energy x-ray absorptiometry scanner for all subjects and visits (Hologic Discovery W; Hologic Inc., Waltham, MA). The intrahepatic lipid (IHL), intramyocellular lipid (IMCL) and extramyocellular lipid (EMCL) content of the proximal part of the right anterior tibial muscle was examined with 1H-MRI spectroscopy using a Signa Excite 1.5T twin speed scanner (GE Medical Systems, Milwaukee, WI) (23). Four muscle spectra were omitted because of technical problems (1 control subject and 3 LBW subjects). The mean full width at half maximum was 12.0 ± 0.1 Hz. The LCModel, version 6.2.1, was used to fit the individual lipid spectra and calculate the lipid/water ratio (24). Brain MRI MRI T1-weighted brain images were acquired with a three-dimensional spoiled gradient recalled scan sequence and an eight-channel, high-resolution brain coil using the following parameters: echo time, 6 ms; repetition time, 22 ms; flip angle, 30°; slice thickness, 1.3 mm; number of slices, 120 to 126; field of view, 26 to 28 cm, and acquisition matrix, 192 × 192. Two brain scans were omitted because of poor image quality (1 LBW subject and 1 control subject). The images were processed with automated segmentation using an ×86-based workstation installed with CentOS, version 6.4, 64-bit operating system, and FreeSurfer, version 5.3 (available at: http://surfer.nmr.mgh.harvard.edu) (25–27). The ventromedial prefrontal cortex (vmPFC) was defined as the sum of the medial and lateral orbitofrontal cortex (28). Longitudinal analysis was performed according to Reuter et al. (29). Physical activity Physical activity was assessed using the self-reported physical activity questionnaire (30), which was further validated in a subgroup of 11 NBW and 27 LBW men using triaxial accelerometry (ActiGraph GT3X). The total number of steps in 24 hours correlated linearly with the total self-reported physical activity questionnaire score (r = 0.54; P < 0.001). Statistical analysis Differences between the LBW and NBW subjects were tested using either an unpaired t test or the nonparametric Wilcoxon (Mann-Whitney) rank sum test according to the distribution. In part 2, group delta values (baseline vs after treatment) were analyzed using unpaired tests. The Pearson correlation test was applied to linear correlations. The STATA®/IC 11.2 statistical software package was used for statistical analysis. P values < 0.05 were considered statistically significant. Results Subject characteristics The LBW subjects were 8.8 ± 1.8 cm shorter and weighed 6.61 ± 2.15 kg less than the NBW subjects (Table 1). No differences were found in the BMI, age, or physical activity between the two groups. The LBW men exhibited a trend toward a greater waist/hip ratio (P = 0.06) and had a statistically significant greater total fat percentage (P = 0.03) and trunk fat/total fat ratio (P < 0.01). MR spectroscopy revealed statistically significantly greater IHL content level in LBW vs NBW men (P < 0.01). In contrast, no differences were found in IMCL or EMCL content. Escitalopram vs placebo in the LBW subjects did not change body composition, IHL content, muscular lipid levels, or physical activity. Table 1. Anthropometrics, Limbic Brain Morphology, and Body Composition for LBW vs NBW Men (Part 1) Variable  NBW (n = 19/20)  LBW (n = 39/40)  P Value  Anthropometric data         Birth weight, g  3775 (3600–3885)a  2350 (2150–2400)a  < 0.001   Term, wk  40 (39–41)a  38 (38–39)a  < 0.001   Age, y  25.0 ± 0.7  24.7 ± 0.5  0.76   Height, cm  185.6 ± 1.6  176.9 ± 1.0  < 0.001   Weight, kg  77.0 ± 1.9  70.4 ± 1.2  < 0.01   BMI, kg/m2  22.3 ± 0.4  22.5 ± 0.3  0.68   Physical activity,b arbitrary units  43.1 (39.5–48.1)a  42.4 (38.4–48.1)a  0.66  Body composition         Waist/hip ratio  0.85 ± 0.01  0.87 ± 0.01  0.06   Total body fat, %c  14.6 (12.9–16.5)a  17.5 (14.7–20.2)a  0.03   Trunk/total fat ratio, arbitrary unitsb  0.44 ± 0.010  0.48 ± 0.008  < 0.01   IMCL content, arbitrary units  0.88 ± 0.11  1.07 ± 0.09  0.21   EMCL content, arbitrary units  2.42 ± 0.17  2.25 ± 0.11  0.38   IHL content, arbitrary units  2.25 ± 0.14  3.00 ± 0.15  < 0.01  Limbic brain morphologyd         Left hippocampus volume, mm3  4369.9 ± 82.1  4384.2 ± 59.8  0.60   Right hippocampus volume, mm3  4338.9 ± 102.1  4444.9 ± 62.7  0.18   Left thalamus volume, mm3  8947.1 ± 255.1  8116.0 ± 111.6  < 0.01   Right thalamus volume, mm3  7923.2 ± 187.1  7671.7 ± 99.2  0.44   Left amygdala volume, mm3  1523.4 ± 34.5  1560.6 ± 26.2  0.21   Right amygdala volume, mm3  1574.1 ± 27.8  1586.3 ± 25.8  0.57   Left vmPFC volume, mm3  14,692.4 ± 362.3  14,252.6 ± 266.0  0.38   Right vmPFC volume, mm3  14,398.9 ± 314.9  14,082.9 ± 210.0  0.52   Left vmPFC surface area, mm2  4822.8 ± 125.9  4450.8 ± 84.6  < 0.01   Right vmPFC surface area, mm2  4742.2 ± 116.8  4462.3 ± 75.5  0.02   Total estimated ICV, mm3  1,712,039 ± 42,296  1,688,664 ± 21,231  0.58  Variable  NBW (n = 19/20)  LBW (n = 39/40)  P Value  Anthropometric data         Birth weight, g  3775 (3600–3885)a  2350 (2150–2400)a  < 0.001   Term, wk  40 (39–41)a  38 (38–39)a  < 0.001   Age, y  25.0 ± 0.7  24.7 ± 0.5  0.76   Height, cm  185.6 ± 1.6  176.9 ± 1.0  < 0.001   Weight, kg  77.0 ± 1.9  70.4 ± 1.2  < 0.01   BMI, kg/m2  22.3 ± 0.4  22.5 ± 0.3  0.68   Physical activity,b arbitrary units  43.1 (39.5–48.1)a  42.4 (38.4–48.1)a  0.66  Body composition         Waist/hip ratio  0.85 ± 0.01  0.87 ± 0.01  0.06   Total body fat, %c  14.6 (12.9–16.5)a  17.5 (14.7–20.2)a  0.03   Trunk/total fat ratio, arbitrary unitsb  0.44 ± 0.010  0.48 ± 0.008  < 0.01   IMCL content, arbitrary units  0.88 ± 0.11  1.07 ± 0.09  0.21   EMCL content, arbitrary units  2.42 ± 0.17  2.25 ± 0.11  0.38   IHL content, arbitrary units  2.25 ± 0.14  3.00 ± 0.15  < 0.01  Limbic brain morphologyd         Left hippocampus volume, mm3  4369.9 ± 82.1  4384.2 ± 59.8  0.60   Right hippocampus volume, mm3  4338.9 ± 102.1  4444.9 ± 62.7  0.18   Left thalamus volume, mm3  8947.1 ± 255.1  8116.0 ± 111.6  < 0.01   Right thalamus volume, mm3  7923.2 ± 187.1  7671.7 ± 99.2  0.44   Left amygdala volume, mm3  1523.4 ± 34.5  1560.6 ± 26.2  0.21   Right amygdala volume, mm3  1574.1 ± 27.8  1586.3 ± 25.8  0.57   Left vmPFC volume, mm3  14,692.4 ± 362.3  14,252.6 ± 266.0  0.38   Right vmPFC volume, mm3  14,398.9 ± 314.9  14,082.9 ± 210.0  0.52   Left vmPFC surface area, mm2  4822.8 ± 125.9  4450.8 ± 84.6  < 0.01   Right vmPFC surface area, mm2  4742.2 ± 116.8  4462.3 ± 75.5  0.02   Total estimated ICV, mm3  1,712,039 ± 42,296  1,688,664 ± 21,231  0.58  Data presented as mean ± standard error of the mean, unless otherwise indicated. Abbreviation: ICV, intracerebral volume. a Non-Gaussian variables presented as median (interquartile range). b Self-reported physical activity questionnaire score. c Estimated from dual energy x-ray absorptiometry scans. d All measures of limbic brain morphology (except for total ICV) were adjusted for total estimated ICV before statistical analyses; p values shown are from unpaired t tests. The found differences in left vmPFC surface area and left thalamus volume remained statistically significant after conservative Bonferroni correction of multiple comparison (data not shown); in contrast, the difference in right vmPFC surface area was no longer statistically significant (P = 0.016; α level = 0.005). View Large LHPA axis characterization In part 1, the 24-hour plasma hormone profiles (Fig. 1A and 1B) revealed a substantial elevation at 5:00 am for plasma levels of free cortisol (39.4 ± 4.0 vs 19.8 ± 3.4 nmol/L; P < 0.001) and ACTH (37.3 ± 4.0 vs 26.3 ± 2.4 pg/mL; P = 0.02) in LBW vs NBW subjects. Although the CBG levels were slightly decreased in LBW vs NBW men (22.9 ± 0.5 vs 24.8 ± 0.6 µg/mL; P = 0.03), the difference for the 5:00-am free cortisol level was primarily caused by a large increase in total cortisol (315.1 ± 22.1 vs 209.9 ± 23.6 nmol/L; P < 0.01). Accordingly, the 24-hour AUC of free cortisol (575.5 ± 24.1 vs 480.5 ± 32.5 nmol/L/h; P = 0.03) was 19.8% ± 8.6% greater in the LBW than in the NBW subjects (Fig. 1C), a difference that predominantly seemed driven by the 5:00 am difference. No statistically significant difference was found in the ACTH AUC over 24 hours (P = 0.75). The plasma levels of free cortisol in our data set were not related to the total or truncal fat content; hence, the differences seen remained statistically significant after adjustment for differences in whole body fat content (data not shown). Figure 1. View largeDownload slide Profiles of 24-hour plasma-free cortisol and ACTH for LBW vs NBW men (part 1). Baseline 24-hour plasma concentration curves of (A) ACTH and (B) free cortisol (NBW indicated by dotted line and white squares; LBW, by solid line and black squares). (C) Total 24-hour AUC of (left) free cortisol and (right) ACTH (NBW, white bars; LBW, black bars). Data presented as mean ± standard error of the mean (n = 20 to 40; *P < 0.05 for LBW vs NBW; **P < 0.001 for LBW vs NBW). Figure 1. View largeDownload slide Profiles of 24-hour plasma-free cortisol and ACTH for LBW vs NBW men (part 1). Baseline 24-hour plasma concentration curves of (A) ACTH and (B) free cortisol (NBW indicated by dotted line and white squares; LBW, by solid line and black squares). (C) Total 24-hour AUC of (left) free cortisol and (right) ACTH (NBW, white bars; LBW, black bars). Data presented as mean ± standard error of the mean (n = 20 to 40; *P < 0.05 for LBW vs NBW; **P < 0.001 for LBW vs NBW). In part 2, treatment of LBW with escitalopram vs placebo normalized the levels of ACTH and cortisol and led to marked reductions in 5:00 am levels of both free cortisol (−21.5 ± 5.1 vs −4.6 ± 6.3 nmol/L; P = 0.04), total cortisol (−107.9 ± 30.3 vs −12.6 ± 35.7 nmol/L; P = 0.04) and ACTH (−20.6 ± 5.7 vs 1.8 ± 3.1 pg/mL; P = 0.002; Table 2). Escitalopram vs placebo did not alter plasma CBG (P = 0.93). The 24-hour AUC of ACTH (−84.6 ± 28.5 vs 5.1 ± 20.6; P = 0.02), but not free cortisol, also decreased statistically significantly with escitalopram vs placebo (−274.5 ± 190.2 vs −111.1 ± 211.6 pg/mL; P = 0.57). Table 2. Changes After 3 Months of Treatment (Escitalopram or Placebo) in LBW Men (Part 2) Variable  Direction of Statistically Significant Differences: LBW vs NBW at Baseline  Δ Values a  Direction of Statistically Significant Changes in LBW With Escitalopram vs Placebo  P Value  LBW, Placebo (n = 20/19)  LBW, Escitalopram (n = 20/19)  Anthropometric data             BMI, kg/m2    0.075 ± 0.12  −0.065 ± 0.13    0.44   Physical activity,b arbitrary units    1.18 ± 1.62  3.03 ± 1.31    0.38  Body composition             Waist/hip ratio    −0.03 (−0.09 to −0.01)c  0.01 (−0.00 to 0.06)c    0.14   Total body fat, %  ↑  0.765 ± 0.24  0.72 ± 0.41    0.93   Trunk/total fat ratio, arbitrary units  ↑  0.002 ± 0.005  0.005 ± 0.006    0.71   IMCL content, arbitrary units    −0.29 ± 0.12  0.08 ± 0.17    0.08   EMCL content, arbitrary units    1.17 ± 0.21  1.46 ± 0.22    0.35   IHL content, arbitrary units  ↑  −0.19 ± 0.17  0.04 ± 0.09    0.24  Limbic brain morphology             Left hippocampus volume, mm3    −28.8 ± 28.2  34.1 ± 23.4    0.09   Right hippocampus volume, mm3    −59.6 ± 29.3  19.5 ± 28.7    0.06   Left thalamus volume, mm3  ↓  −58.5 ± 44.4  −23.6 ± 71.8    0.49   Right thalamus volume, mm3    21.5 ± 64.8  −3.4 ± 54.9    0.77   Left amygdala volume, mm3    −21.6 ± 16.4  14.2 ± 16.4    0.13   Right amygdala volume, mm3    16.6 ± 18.5  7.6 ± 15.8    0.71   Left vmPFC volume, mm3    −28.3 ± 108.5  80.2 ± 181.0    0.61   Right vmPFC volume, mm3    −176.4 ± 86.5  105.5 ± 188.1    0.18   Left vmPFC surface area, mm2  ↓  −21.4 ± 22.6  24.1 ± 19.2    0.14   Right vmPFC surface area, mm2  ↓  −14.2 ± 19.0  1.2 ± 36.3    0.71   Total estimated ICV, mm3    —  —    —  Liver glucose metabolism             EGPbasal, Δ in %    4.89 ± 4.47  3.55 ± 4.39    0.83   EGPsubmax, Δ in %    32.12 ± 21.10  20.00 ± 20.55    0.68   HSI, Δ in %  ↓  −11.96 ± 5.89  −2.86 ± 5.36    0.26  Peripheral glucose metabolism             Rsubmax, Δ in %  ↓  3.7 ± 5.2  23.9 ± 8.4  ↑  0.04   Rdmax, Δ in %  ↓  −3.9 ± 3.9  7.3 ± 4.9    0.08   Rdsubmax/insulin ratio, Δ in %  ↓  −3.4 ± 4.9  20.1 ± 9.1  ↑  0.03   Rdmax/insulin ratio, Δ in %  ↓  −6.5 ± 5.8  6.9 ± 6.3    0.12  LHPA axis             Free cortisol AUC-24 h, nmol/L/h  ↑  −49.1 ± 50.8  −84.6 ± 28.9    0.55   ACTH AUC 24-h, pg/mL/h    5.1 ± 20.6  −84.6 ± 28.5  ↓  0.01   5 am ACTH concentration, pg/mL  ↑  1.8 ± 3.1  −20.6 ± 5.7  ↓  0.002   5 am free cortisol concentration, nmol/L  ↑  −4.6 ± 6.3  −21.5 ± 5.1  ↓  0.04  Variable  Direction of Statistically Significant Differences: LBW vs NBW at Baseline  Δ Values a  Direction of Statistically Significant Changes in LBW With Escitalopram vs Placebo  P Value  LBW, Placebo (n = 20/19)  LBW, Escitalopram (n = 20/19)  Anthropometric data             BMI, kg/m2    0.075 ± 0.12  −0.065 ± 0.13    0.44   Physical activity,b arbitrary units    1.18 ± 1.62  3.03 ± 1.31    0.38  Body composition             Waist/hip ratio    −0.03 (−0.09 to −0.01)c  0.01 (−0.00 to 0.06)c    0.14   Total body fat, %  ↑  0.765 ± 0.24  0.72 ± 0.41    0.93   Trunk/total fat ratio, arbitrary units  ↑  0.002 ± 0.005  0.005 ± 0.006    0.71   IMCL content, arbitrary units    −0.29 ± 0.12  0.08 ± 0.17    0.08   EMCL content, arbitrary units    1.17 ± 0.21  1.46 ± 0.22    0.35   IHL content, arbitrary units  ↑  −0.19 ± 0.17  0.04 ± 0.09    0.24  Limbic brain morphology             Left hippocampus volume, mm3    −28.8 ± 28.2  34.1 ± 23.4    0.09   Right hippocampus volume, mm3    −59.6 ± 29.3  19.5 ± 28.7    0.06   Left thalamus volume, mm3  ↓  −58.5 ± 44.4  −23.6 ± 71.8    0.49   Right thalamus volume, mm3    21.5 ± 64.8  −3.4 ± 54.9    0.77   Left amygdala volume, mm3    −21.6 ± 16.4  14.2 ± 16.4    0.13   Right amygdala volume, mm3    16.6 ± 18.5  7.6 ± 15.8    0.71   Left vmPFC volume, mm3    −28.3 ± 108.5  80.2 ± 181.0    0.61   Right vmPFC volume, mm3    −176.4 ± 86.5  105.5 ± 188.1    0.18   Left vmPFC surface area, mm2  ↓  −21.4 ± 22.6  24.1 ± 19.2    0.14   Right vmPFC surface area, mm2  ↓  −14.2 ± 19.0  1.2 ± 36.3    0.71   Total estimated ICV, mm3    —  —    —  Liver glucose metabolism             EGPbasal, Δ in %    4.89 ± 4.47  3.55 ± 4.39    0.83   EGPsubmax, Δ in %    32.12 ± 21.10  20.00 ± 20.55    0.68   HSI, Δ in %  ↓  −11.96 ± 5.89  −2.86 ± 5.36    0.26  Peripheral glucose metabolism             Rsubmax, Δ in %  ↓  3.7 ± 5.2  23.9 ± 8.4  ↑  0.04   Rdmax, Δ in %  ↓  −3.9 ± 3.9  7.3 ± 4.9    0.08   Rdsubmax/insulin ratio, Δ in %  ↓  −3.4 ± 4.9  20.1 ± 9.1  ↑  0.03   Rdmax/insulin ratio, Δ in %  ↓  −6.5 ± 5.8  6.9 ± 6.3    0.12  LHPA axis             Free cortisol AUC-24 h, nmol/L/h  ↑  −49.1 ± 50.8  −84.6 ± 28.9    0.55   ACTH AUC 24-h, pg/mL/h    5.1 ± 20.6  −84.6 ± 28.5  ↓  0.01   5 am ACTH concentration, pg/mL  ↑  1.8 ± 3.1  −20.6 ± 5.7  ↓  0.002   5 am free cortisol concentration, nmol/L  ↑  −4.6 ± 6.3  −21.5 ± 5.1  ↓  0.04  Data presented as mean ± standard error of the mean, unless otherwise indicated. Abbreviation: basal, fasting conditions; ICV, intracerebral volume; max, maximal insulin stimulation (e.g., 1.0 mU/kg/min); submax, submaximal insulin stimulation (e.g., 0.3 mU/kg/min). a Absolute changes, unless otherwise indicated. b Self-reported physical activity questionnaire score. c Non-Gaussian variables presented as median (interquartile range). View Large Assessments of insulin sensitivity In part 1, the LBW and NBW men displayed comparable fasting plasma glucose levels (5.03 ± 0.08 vs 5.09 ± 0.05 mM; P = 0.47), although the LBW subjects exhibited statistically significant greater fasting insulin levels (36.4 ± 2.0 vs 27.0 ± 1.7 pmol/L; P < 0.001). The average steady-state insulin levels during the clamp were also slightly greater in the LBW group at both submaximal (130.4 ± 3.6 pM vs 105.5 ± 4.7 pM; P < 0.001) and maximal insulin stimulation (402.8 ± 9.5 pmol/L vs 335.9 ± 12.1 pmol/L; P < 0.001; Fig. 2). Because the insulin infusions were performed according to body weight; this difference might be ascribed, in part, to relatively more fat vs lean mass in the LBW vs NBW subjects. Nevertheless, despite higher steady-state insulin levels at both stage 1 (∼23.5%) and stage 2 (∼19.9%) in the LBW vs NBW groups, the glucose infusion rates at each stage were lower in the LBW vs NBW men (P < 0.05 and P < 0.01, respectively; Fig. 2), and the Rd estimates were significantly lower statistically in the LBW men (Rdsubmax, 3.51 ± 0.14 vs 4.17 ± 0.25; P = 0.01; Rdmax, 9.21 ± 0.30 vs 10.47 ± 0.31; P = 0.01; Fig. 3A). Furthermore, when the rates of glucose disposal were adjusted for higher steady-state insulin levels (Rd/insulin ratio; Fig. 3A), the differences seen in LBW vs NBW groups became even more pronounced. Together, these data strongly indicate the presence of peripheral insulin resistance in the LBW phenotype. Figure 2. View largeDownload slide Clamp rates of glucose infusion (GIR) and plasma insulin (p-insulin) concentrations for LBW vs NBW men (part 1). Primary y-axis shows GIRs to maintain euglycemia during stage 1 and 2, and secondary y-axis shows plasma insulin concentrations during basal and insulin-stimulated conditions. The last 30 minutes of each experimental period or stage were used to estimate the average steady-state plasma insulin concentration and GIR, respectively. Data presented as mean ± standard error of the mean (n = 20 to 40; *P < 0.01, LBW vs NBW; **P < 0.001, LBW vs NBW; ***P < 0.05, LBW vs NBW). Figure 2. View largeDownload slide Clamp rates of glucose infusion (GIR) and plasma insulin (p-insulin) concentrations for LBW vs NBW men (part 1). Primary y-axis shows GIRs to maintain euglycemia during stage 1 and 2, and secondary y-axis shows plasma insulin concentrations during basal and insulin-stimulated conditions. The last 30 minutes of each experimental period or stage were used to estimate the average steady-state plasma insulin concentration and GIR, respectively. Data presented as mean ± standard error of the mean (n = 20 to 40; *P < 0.01, LBW vs NBW; **P < 0.001, LBW vs NBW; ***P < 0.05, LBW vs NBW). Figure 3. View largeDownload slide Estimated rates of glucose turnover and insulin sensitivity indexes for LBW vs NBW men (part 1). (A) Basal and insulin-stimulated rates of peripheral (e.g., Rd) and (B) EGP and related estimated insulin sensitivity indexes [e.g., Rd/insulin ratios (A) and HSI (B)]. Data presented as mean ± standard error of the mean (n = 20 to 40; *P = 0.01, LBW vs NBW; **P < 0.001, LBW vs NBW; ***P < 0.0001, LBW vs NBW; ****P < 0.03, LBW vs NBW). Basal, fasting conditions; max, maximal insulin stimulation (e.g., 1.0 mU/kg/min); submax, submaximal insulin stimulation (e.g., 0.3 mU/kg/min). Figure 3. View largeDownload slide Estimated rates of glucose turnover and insulin sensitivity indexes for LBW vs NBW men (part 1). (A) Basal and insulin-stimulated rates of peripheral (e.g., Rd) and (B) EGP and related estimated insulin sensitivity indexes [e.g., Rd/insulin ratios (A) and HSI (B)]. Data presented as mean ± standard error of the mean (n = 20 to 40; *P = 0.01, LBW vs NBW; **P < 0.001, LBW vs NBW; ***P < 0.0001, LBW vs NBW; ****P < 0.03, LBW vs NBW). Basal, fasting conditions; max, maximal insulin stimulation (e.g., 1.0 mU/kg/min); submax, submaximal insulin stimulation (e.g., 0.3 mU/kg/min). Regarding the rates of EGP, we were unable to detect any differences between the LBW vs NBW groups, neither in the basal nor in the submaximally insulin-stimulated state (Fig. 3B), and no difference was found in the insulin-mediated relative suppression of EGP (data not shown). However, the HSI was significantly lower statistically in the LBW vs NBW men [5.60 ± 0.37 vs 7.06 ± 0.60 (mg/kg/min) × pmol/L × 10−4; P = 0.03], suggesting that hepatic insulin resistance might have been overlooked in our clamp experiment owing to the higher insulin levels. In addition to these findings, the 20 LBW subjects with 5:00 am free cortisol greater than the median (60.2 ± 3.6 nmol/L) had a statistically significant lower Rd at submaximal insulin stimulation (3.16 ± 0.14 vs 3.85 ± 0.21 mg/kg/min; P < 0.01) and a lower basal HSI (4.53 ± 0.36 vs 6.67 ± 0.55 [10,000/(mg/kg/min) × pM]; P = 0.002) compared with the 20 LBW subjects with a 5:00 am free cortisol less than the median (19.5 ± 2.8 nmol/L), suggesting even more pronounced insulin resistance in the LBW subjects with the highest free cortisol levels. No differences were found in age, BMI, truncal fat percentage or physical activity between these two subgroups of LBW individuals (data not shown). In the clamps performed in part 2 on LBW subjects exposed to either escitalopram or placebo, respectively, the insulin levels were the same (data not shown). Treatment with escitalopram vs placebo significantly increased (with statistical significance) the Rdsubmax and Rdsubmax/insulin ratio in LBW subjects at submaximal insulin stimulation by ∼24% and ∼20% (23.9% ± 8.4% vs 3.7% ± 5.2%, P = 0.04; 20.1% ± 9.1% vs −3.4% ± 4.9%, P = 0.03; Table 2). Although a trend was detected, no statistically significant effect of escitalopram vs placebo on the Rdmax or Rdmax/insulin ratio was obtained during maximal insulin stimulation (7.3% ± 4.9% vs −3.9% ± 3.9%, P = 0.08; 6.9% ± 6.3% vs −6.5% ± 5.8%, P = 0.12). Limbic brain morphology LBW vs NBW subjects (part 1) were characterized by marked reductions in the volume of the left thalamus (8.2% ± 2.9%; P < 0.01) and surface areas of both left (6.8% ± 2.3%; P < 0.01) and right (4.9% ± 2.0%; P = 0.02) vmPFC (Table 1). These specific limbic measures further correlated inversely with the plasma levels of 5:00 am free cortisol (Fig. 4A–4C; left thalamus volume, r = −0.37, P = 0.02; right vmPFC surface area, r = −0.39, P = 0.02; left vmPFC surface area, r = −0.41; P = 0.01). The correlations with 24-hour free cortisol AUC were much weaker and statistically nonsignificant (data not shown). However, no differences were found between the LBW and NBW subjects in the volumes of the right thalamus, amygdalae, or hippocampi (Table 1). Figure 4. View largeDownload slide Correlations of 5:00 am free cortisol to key limbic brain measures for LBW men (part 1). Pearson linear correlations between 5:00 am free cortisol and key limbic measures in LBW subjects. Scatter plots for 5:00 am free cortisol vs surface areas of (A) left and (B) right vmPFC. (C) Scatter plot for 5:00 am free cortisol vs left thalamus volume. The related correlation coefficients (r) and p values are indicated in the upper right corner of each panel; n = 39 for each plot. Figure 4. View largeDownload slide Correlations of 5:00 am free cortisol to key limbic brain measures for LBW men (part 1). Pearson linear correlations between 5:00 am free cortisol and key limbic measures in LBW subjects. Scatter plots for 5:00 am free cortisol vs surface areas of (A) left and (B) right vmPFC. (C) Scatter plot for 5:00 am free cortisol vs left thalamus volume. The related correlation coefficients (r) and p values are indicated in the upper right corner of each panel; n = 39 for each plot. In the LBW subjects exposed for 3 months to escitalopram vs placebo (part 2), longitudinal analyses could not reveal any statistically significant changes in the cortical volumes, subcortical volumes, or surface areas (Table 2). Discussion In part 1 of our investigation (LBW phenotype study), we found that young LBW men compared with the matched NBW controls exhibited an ∼20% increment in 24-hour free cortisol AUC, primarily driven by a prominent increase in the early morning/late night (5:00 am) plasma-free cortisol with an equivalent increase in plasma ACTH levels. In humans, the limbic brain structures such as the hippocampus, amygdala, and vmPFC and the dorsal, anteroventral, and paraventricular nuclei of the thalamus are known to be involved in the intrinsic regulation of both circadian rhythms and the reactivity of the LHPA axis (11, 12). Moreover, the thalamus seems to play an important role in the adaptive stress response (31), at same time, modulating both diurnal and stress-related hypothalamic release of corticotrophin-releasing hormone (32). We detected reductions in the volume of the left thalamus and the surface areas of the left and right vmPFC. These specific limbic measures further correlated inversely with elevated 5:00 am free-plasma cortisol levels but not with the 24-hour cortisol AUC. Hence, it appears that the limbic changes were primarily linked to late nighttime hyperactivity of the LHPA axis rather than 24-hour LHPA axis hyperactivity. Furthermore, as glucocorticoids stimulate lipolysis, our present finding of a specific increase in late night cortisol could be in line with a report by Reuter et al. (29) demonstrating diurnal differences in the rates of lipolysis, with greater rates during the night than during the day in young LBW men. Moreover, although LHPA axis disturbances previously have been reported in LBW subjects (4, 5, 18), to the best of our knowledge, the present study is the first to provide evidence that LHPA axis hyperactivity in LBW individual might be attributable to structural anomalies within key limbic organs such as the thalamus and vmPFC. Glucocorticoids are known to cause insulin resistance in skeletal muscle tissue (7), and the present LBW subjects displayed peripheral insulin resistance with lower rates of glucose disappearance (Rd) at both submaximal and maximal insulin stimulation. Moreover, insulin resistance was substantially worse in those LBW men within the upper half of the range of 5:00 am free cortisol levels. Hence, based on our findings from part 1, it would be tempting for us to hypothesize that a causal relationship might exist between LHPA axis hyperactivity and insulin resistance in LBW individuals. Although lipid accumulation in skeletal muscle tissues has been associated with insulin resistance (33), the insulin-resistant LBW vs NBW men did not display any increases in either IMCL or EMCL content. A similar disassociation between muscle fat and muscle insulin action has been reported by Dufour and Petersen (34) in LBW humans. A possible explanation for this might be that glucocorticoids seem to both stimulate intramyocellular lipolysis and inhibit lipogenesis (35), thereby potentially minimizing lipid accumulation in muscle tissue despite increased whole body fat content. Both basal and insulin-stimulated EGP rates and the relative degree of insulin-mediated EGP suppression were comparable between LBW and NBW subjects. Nevertheless, it is likely that the hepatic insulin sensitivity was overestimated in the LBW men owing to the greater insulin levels during the clamp. Hence, we decided, in addition, to calculate the basal HSI (22), a validated alternative measure of hepatic insulin action. The basal HSI was substantially lower in the LBW than in the NBW men. In addition, we found increased levels of IHL content in the livers of the LBW compared with the NBW subjects, a finding that consistently has been linked to hepatic insulin resistance (36). Hence, overall, the presented data could still be in support of previous findings of liver insulin resistance in humans with a LBW (37). In part 2 of our investigation (the LBW trial), the LBW subjects were treated with escitalopram or placebo for 3 months in a randomized and double-blind manner. The intervention with escitalopram was capable of normalizing the plasma levels of 5:00 am free cortisol and ACTH and, at the same time, vastly improving peripheral insulin sensitivity by ∼24% as measured at submaximal insulin stimulation. However, insulin responsiveness at maximal insulin stimulation was only slightly improved without reaching statistical significance. Furthermore, these improvements in insulin action and LHPA axis homeostasis were not related to changes in body composition or physical activity. Similar beneficial effects with SSRI compounds on the LHPA axis and insulin action have been reported by us in a previous study performed on LBW rats (18) and by others of patients with major depressive disorder (15). In addition, Breum et al. (38) demonstrated that SSRI treatment of obese patients with T2D improved glycemic control independently of weight loss. The investigators speculated that the insulin-sensitizing effect might have resulted from upregulation of glycogen synthase activity in skeletal muscle. This is an interesting view, because it has been shown that cortisol exposure of skeletal muscle tissue leads to insulin resistance, accompanied by a marked reduction in glycogen synthase activity (39). Thus, it can be speculated further that the beneficial effects of SSRI treatment in the study by Breum et al. (38) might have resulted from reduced cortisol exposure, although this was not measured. Although a direct beneficial effect of escitalopram on muscle tissue insulin action could not be ruled out, data are available that speak against this notion. First, a previous study reported that serotonin receptor activation in human skeletal muscle had no effect on insulin-mediated glucose uptake (40). Second, in pilot studies, we have previously seen that the effects of escitalopram regarding insulin action were highly phenotype dependent, with a negative effect on glucose metabolism (and the LHPA axis) in NBW rats, despite positive effects in LBW rats (18). Together, these data speak against the idea that escitalopram exerts a direct insulin-sensitizing effect on the peripheral tissues via local serotonin receptor activation. Escitalopram exposure did not completely restore insulin action in LBW subjects. However, escitalopram exposure also failed to improve the adverse body composition and fatty liver seen in LBW subjects. These remaining negative metabolic defects seen after normalization of cortisol levels might contribute to the residual insulin resistance. Nevertheless, when the positive metabolic and neuroendocrine effects of escitalopram seen in the LBW trial (part 2) are considered with the findings of insulin resistance and increased free cortisol levels in the LBW phenotype study (part 1), our results indicate that escitalopram improved insulin action in LBW subjects via a reduction in cortisol. Hence, the findings from part 2 provide further support for our hypothesis that hypercortisolemia might play an etiological role in the development of insulin resistance in LBW subjects. In conclusion, we have presented evidence that young men born with LBW exhibit morphological changes within key components of the limbic system that are linked to increased late night or early morning secretion of ACTH and cortisol. Furthermore, the presented data suggest that these neuroendocrine changes might, at least in part, be causally associated to the development of peripheral insulin resistance. Finally, the present investigation has shown that an SSRI compound can be used to reverse LHPA axis hyperactivity and ameliorate peripheral insulin resistance in LBW humans. Hence, compounds ameliorating insulin resistance by targeting defects in the LHPA axis could be clinically useful for reducing the risk of T2D in individuals born with a LBW. Abbreviations: ACTH adrenocorticotropic hormone AUC area under the curve BMI body mass index CBG corticosteroid-binding globulin EGP endogenous glucose production GCP Good Clinical Practice HSI hepatic insulin sensitivity index LBW low birthweight LHPA limbic–hypothalamic–pituitary–adrenal MRI magnetic resonance imaging NBW normal birthweight Rd rate of glucose disappearance Rdmax maximal insulin-stimulated glucose turnover Rdsubmax submaximal insulin-stimulated glucose turnover SSRI selective serotonin reuptake inhibitor T2D type 2 diabetes vmPFC ventromedial prefrontal cortex. Acknowledgments We thank the study participants and the laboratory technicians Eva Schriver, Karen Mathiassen, Lone Kvist, Dorthe Wulf, and Joan Hansen for their contributions. Special thanks to Annette Mengel for help with the hyperinsulinemic euglycemic clamps. Thanks also to Professor Kitt Falk Petersen, Department of Internal Medicine, Yale University School of Medicine, New Haven, Connecticut, for inspiring scientific discussions. Financial Support: The present study was funded by an independent investigator grant from H. Lundbeck A/S, a PhD grant from Aarhus University, the Augustinusfonden and Aase and Ejnar Danielsen Fond. H. Lundbeck A/S supplied the trial medication. Clinical Trial Information: ClinicalTrials.gov identifier, NCT00971815 (registered 4 September 2009). Disclosure Summary: The study was investigator initiated, and, apart from the independent investigator grant from H. Lundbeck A/S, the authors have nothing to disclose. References 1. Harder T, Rodekamp E, Schellong K, Dudenhausen JW, Plagemann A. Birth weight and subsequent risk of type 2 diabetes: a meta-analysis. Am J Epidemiol . 2007; 165( 8): 849– 857. Google Scholar CrossRef Search ADS PubMed  2. Hales CN, Barker DJ. The thrifty phenotype hypothesis. Br Med Bull . 2001; 60( 1): 5– 20. Google Scholar CrossRef Search ADS PubMed  3. Meher S, Hernandez-Andrade E, Basheer SN, Lees C. Impact of cerebral redistribution on neurodevelopmental outcome in small-for-gestational-age or growth-restricted babies: a systematic review. Ultrasound Obstet Gynecol . 2015; 46( 4): 398– 404. Google Scholar CrossRef Search ADS PubMed  4. Phillips DI, Walker BR, Reynolds RM, Flanagan DE, Wood PJ, Osmond C, Barker DJ, Whorwood CB. Low birth weight predicts elevated plasma cortisol concentrations in adults from 3 populations. Hypertension . 2000; 35( 6): 1301– 1306. Google Scholar CrossRef Search ADS PubMed  5. Phillips DI, Barker DJ, Fall CH, Seckl JR, Whorwood CB, Wood PJ, Walker BR. Elevated plasma cortisol concentrations: a link between low birth weight and the insulin resistance syndrome? J Clin Endocrinol Metab . 1998; 83( 3): 757– 760. Google Scholar PubMed  6. Kajantie E, Feldt K, Räikkönen K, Phillips DI, Osmond C, Heinonen K, Pesonen AK, Andersson S, Barker DJ, Eriksson JG. Body size at birth predicts hypothalamic-pituitary-adrenal axis response to psychosocial stress at age 60 to 70 years. J Clin Endocrinol Metab . 2007; 92( 11): 4094– 4100. Google Scholar CrossRef Search ADS PubMed  7. Weinstein SP, Paquin T, Pritsker A, Haber RS. Glucocorticoid-induced insulin resistance: dexamethasone inhibits the activation of glucose transport in rat skeletal muscle by both insulin- and non-insulin-related stimuli. Diabetes . 1995; 44( 4): 441– 445. Google Scholar CrossRef Search ADS PubMed  8. Lowe J, Duvall SW, MacLean PC, Caprihan A, Ohls R, Qualls C, Phillips J. Comparison of structural magnetic resonance imaging and development in toddlers born very low birth weight and full-term. J Child Neurol . 2011; 26( 5): 586– 592. Google Scholar CrossRef Search ADS PubMed  9. Martinussen M, Flanders DW, Fischl B, Busa E, Løhaugen GC, Skranes J, Vangberg TR, Brubakk AM, Haraldseth O, Dale AM. Segmental brain volumes and cognitive and perceptual correlates in 15-year-old adolescents with low birth weight. J Pediatr . 2009; 155( 6): 848– 853.e1. Google Scholar CrossRef Search ADS PubMed  10. Skranes J, Løhaugen GC, Martinussen M, Håberg A, Brubakk AM, Dale AM. Cortical surface area and IQ in very-low-birth-weight (VLBW) young adults. Cortex . 2013; 49( 8): 2264– 2271. Google Scholar CrossRef Search ADS PubMed  11. Ahima R, Krozowski Z, Harlan R. Type I corticosteroid receptor-like immunoreactivity in the rat CNS: distribution and regulation by corticosteroids. J Comp Neurol . 1991; 313( 3): 522– 538. Google Scholar CrossRef Search ADS PubMed  12. Morimoto M, Morita N, Ozawa H, Yokoyama K, Kawata M. Distribution of glucocorticoid receptor immunoreactivity and mRNA in the rat brain: an immunohistochemical and in situ hybridization study. Neurosci Res . 1996; 26( 3): 235– 269. Google Scholar CrossRef Search ADS PubMed  13. Piwowarska J, Chimiak A, Matsumoto H, Dziklińska A, Radziwoń-Zaleska M, Szelenberger W, Pachecka J. Serum cortisol concentration in patients with major depression after treatment with fluoxetine. Psychiatry Res . 2012; 198( 3): 407– 411. Google Scholar CrossRef Search ADS PubMed  14. Nikisch G, Mathé AA, Czernik A, Thiele J, Bohner J, Eap CB, Agren H, Baumann P. Long-term citalopram administration reduces responsiveness of HPA axis in patients with major depression: relationship with S-citalopram concentrations in plasma and cerebrospinal fluid (CSF) and clinical response. Psychopharmacology (Berl) . 2005; 181( 4): 751– 760. Google Scholar CrossRef Search ADS PubMed  15. Weber-Hamann B, Gilles M, Lederbogen F, Heuser I, Deuschle M. Improved insulin sensitivity in 80 nondiabetic patients with MDD after clinical remission in a double-blind, randomized trial of amitriptyline and paroxetine. J Clin Psychiatry . 2006; 67( 12): 1856– 1861. Google Scholar CrossRef Search ADS PubMed  16. Lustman PJ, Freedland KE, Griffith LS, Clouse RE. Fluoxetine for depression in diabetes: a randomized double-blind placebo-controlled trial. Diabetes Care . 2000; 23( 5): 618– 623. Google Scholar CrossRef Search ADS PubMed  17. Buhl ES, Neschen S, Yonemitsu S, Rossbacher J, Zhang D, Morino K, Flyvbjerg A, Perret P, Samuel V, Kim J, Cline GW, Petersen KF. Increased hypothalamic-pituitary-adrenal axis activity and hepatic insulin resistance in low-birth-weight rats. Am J Physiol Endocrinol Metab . 2007; 293( 5): E1451– E1458. Google Scholar CrossRef Search ADS PubMed  18. Buhl ES, Jensen TK, Jessen N, Elfving B, Buhl CS, Kristiansen SB, Pold R, Solskov L, Schmitz O, Wegener G, Lund S, Petersen KF. Treatment with an SSRI antidepressant restores hippocampo-hypothalamic corticosteroid feedback and reverses insulin resistance in low-birth-weight rats. Am J Physiol Endocrinol Metab . 2010; 298( 5): E920– E929. Google Scholar CrossRef Search ADS PubMed  19. Coolens JL, Van Baelen H, Heyns W. Clinical use of unbound plasma cortisol as calculated from total cortisol and corticosteroid-binding globulin. J Steroid Biochem . 1987; 26( 2): 197– 202. Google Scholar CrossRef Search ADS PubMed  20. Fall CH, Dennison E, Cooper C, Pringle J, Kellingray SD, Hindmarsh P. Does birth weight predict adult serum cortisol concentrations? Twenty-four-hour profiles in the United kingdom 1920-1930 Hertfordshire Birth Cohort. J Clin Endocrinol Metab . 2002; 87( 5): 2001– 2007. Google Scholar PubMed  21. Steele R. Influences of glucose loading and of injected insulin on hepatic glucose output. Ann N Y Acad Sci . 1959; 82( 2): 420– 430. Google Scholar CrossRef Search ADS PubMed  22. Miyazaki Y, Glass L, Triplitt C, Wajcberg E, Mandarino LJ, DeFronzo RA. Abdominal fat distribution and peripheral and hepatic insulin resistance in type 2 diabetes mellitus. Am J Physiol Endocrinol Metab . 2002; 283( 6): E1135– E1143. Google Scholar CrossRef Search ADS PubMed  23. Wamberg L, Kampmann U, Stødkilde-Jørgensen H, Rejnmark L, Pedersen SB, Richelsen B. Effects of vitamin D supplementation on body fat accumulation, inflammation, and metabolic risk factors in obese adults with low vitamin D levels—results from a randomized trial. Eur J Intern Med . 2013; 24( 7): 644– 649. Google Scholar CrossRef Search ADS PubMed  24. Provencher SW. Estimation of metabolite concentrations from localized in vivo proton NMR spectra. Magn Reson Med . 1993; 30( 6): 672– 679. Google Scholar CrossRef Search ADS PubMed  25. Morey RA, Petty CM, Xu Y, Hayes JP, Wagner HR II, Lewis DV, LaBar KS, Styner M, McCarthy G. A comparison of automated segmentation and manual tracing for quantifying hippocampal and amygdala volumes. Neuroimage . 2009; 45( 3): 855– 866. Google Scholar CrossRef Search ADS PubMed  26. Keller SS, Gerdes JS, Mohammadi S, Kellinghaus C, Kugel H, Deppe K, Ringelstein EB, Evers S, Schwindt W, Deppe M. Volume estimation of the thalamus using FreeSurfer and stereology: consistency between methods. Neuroinformatics . 2012; 10( 4): 341– 350. Google Scholar CrossRef Search ADS PubMed  27. Dale AM, Fischl B, Sereno MI. Cortical surface-based analysis. I. Segmentation and surface reconstruction. Neuroimage . 1999; 9( 2): 179– 194. Google Scholar CrossRef Search ADS PubMed  28. Boes AD, Bechara A, Tranel D, Anderson SW, Richman L, Nopoulos P. Right ventromedial prefrontal cortex: a neuroanatomical correlate of impulse control in boys. Soc Cogn Affect Neurosci . 2009; 4( 1): 1– 9. Google Scholar CrossRef Search ADS PubMed  29. Reuter M, Schmansky NJ, Rosas HD, Fischl B. Within-subject template estimation for unbiased longitudinal image analysis. Neuroimage . 2012; 61( 4): 1402– 1418. Google Scholar CrossRef Search ADS PubMed  30. Aadahl M, Jørgensen T. Validation of a new self-report instrument for measuring physical activity. Med Sci Sports Exerc . 2003; 35( 7): 1196– 1202. Google Scholar CrossRef Search ADS PubMed  31. Jaferi A, Bhatnagar S. Corticosterone can act at the posterior paraventricular thalamus to inhibit hypothalamic-pituitary-adrenal activity in animals that habituate to repeated stress. Endocrinology . 2006; 147( 10): 4917– 4930. Google Scholar CrossRef Search ADS PubMed  32. Hsu DT, Bakshi VP, Roseboom PH, Kalin NH. Diurnal changes in corticotropin-releasing hormone messenger RNA in the rat thalamus. Neurosci Lett . 2003; 338( 1): 33– 36. Google Scholar CrossRef Search ADS PubMed  33. Pan DA, Lillioja S, Kriketos AD, Milner MR, Baur LA, Bogardus C, Jenkins AB, Storlien LH. Skeletal muscle triglyceride levels are inversely related to insulin action. Diabetes . 1997; 46( 6): 983– 988. Google Scholar CrossRef Search ADS PubMed  34. Dufour S, Petersen KF. Disassociation of liver and muscle insulin resistance from ectopic lipid accumulation in low-birth-weight individuals. J Clin Endocrinol Metab . 2011; 96( 12): 3873– 3880. Google Scholar CrossRef Search ADS PubMed  35. Morgan SA, Gathercole LL, Simonet C, Hassan-Smith ZK, Bujalska I, Guest P, Abrahams L, Smith DM, Stewart PM, Lavery GG, Tomlinson JW. Regulation of lipid metabolism by glucocorticoids and 11β-HSD1 in skeletal muscle. Endocrinology . 2013; 154( 7): 2374– 2384. Google Scholar CrossRef Search ADS PubMed  36. Petersen MC, Shulman GI. Roles of diacylglycerols and ceramides in hepatic insulin resistance. Trends Pharmacol Sci.  2017; 38: 649– 665. Google Scholar CrossRef Search ADS PubMed  37. Brøns C, Jensen CB, Storgaard H, Alibegovic A, Jacobsen S, Nilsson E, Astrup A, Quistorff B, Vaag A. Mitochondrial function in skeletal muscle is normal and unrelated to insulin action in young men born with low birth weight. J Clin Endocrinol Metab . 2008; 93( 10): 3885– 3892. Google Scholar CrossRef Search ADS PubMed  38. Breum L, Bjerre U, Bak JF, Jacobsen S, Astrup A. Long-term effects of fluoxetine on glycemic control in obese patients with non-insulin-dependent diabetes mellitus or glucose intolerance: influence on muscle glycogen synthase and insulin receptor kinase activity. Metabolism . 1995; 44( 12): 1570– 1576. Google Scholar CrossRef Search ADS PubMed  39. Holmäng A, Björntorp P. The effects of cortisol on insulin sensitivity in muscle. Acta Physiol Scand . 1992; 144( 4): 425– 431. Google Scholar CrossRef Search ADS PubMed  40. Hajduch E, Rencurel F, Balendran A, Batty IH, Downes CP, Hundal HS. Serotonin (5-Hydroxytryptamine), a novel regulator of glucose transport in rat skeletal muscle. J Biol Chem . 1999; 274( 19): 13563– 13568. Google Scholar CrossRef Search ADS PubMed  Copyright © 2018 Endocrine Society http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Clinical Endocrinology and Metabolism Oxford University Press

Escitalopram Ameliorates Hypercortisolemia and Insulin Resistance in Low Birth Weight Men With Limbic Brain Alterations

Loading next page...
 
/lp/ou_press/escitalopram-ameliorates-hypercortisolemia-and-insulin-resistance-in-50zkWVS13x
Publisher
Endocrine Society
Copyright
Copyright © 2018 Endocrine Society
ISSN
0021-972X
eISSN
1945-7197
D.O.I.
10.1210/jc.2017-01438
Publisher site
See Article on Publisher Site

Abstract

Abstract Context Low birth weight (LBW; <2500 g) is linked to the development of insulin resistance and limbic-hypothalamic-pituitary-adrenal (LHPA) axis hyperactivity. Objective Our first aim was to study insulin action, LHPA axis function, and limbic brain structures in young, healthy LBW men vs normal birthweight (NBW) controls (part 1). Our second aim was to investigate the effects of escitalopram vs placebo in LBW men in the LHPA axis and insulin sensitivity (part 2). Design Setting, Participants, and Intervention The maximal (Rdmax) and submaximal (Rdsubmax) rates of insulin-stimulated glucose turnover, LHPA axis, and brain morphology were examined in 40 LBW men and 20 matched NBW men using two-stage hyperinsulinemic euglycemic clamp, 24-hour hormone plasma profiles, and magnetic resonance imaging. Subsequently, all LBW subjects underwent randomized and double-blind treatment with escitalopram 20 mg/d or placebo for 3 months followed by a complete reexamination. Main Outcome Measures (Part 2) Changes in Rdmax/Rdsubmax and plasma-free cortisol 24-hour area under the curve. Results In LBW vs NBW, Rdsubmax and Rdmax were ∼16% (P = 0.01) and ∼12% (P = 0.01) lower, respectively, and 24-hour free cortisol levels were ∼20% higher (P = 0.02), primarily driven by a ∼99% increase at 05:00 am (P < 0.001). Furthermore, these changes were related to structural alterations within left thalamus and ventromedial prefrontal cortex. However, in LBW men, exposure to escitalopram normalized the free cortisol levels and improved the Rdsubmax by ∼24% (P = 0.04) compared with placebo. Conclusions LBW vs NBW displayed alterations in key brain structures modulating the LHPA axis, elevated free cortisol levels, and insulin resistance. Escitalopram administration ameliorated these defects, suggesting a potential for LHPA axis modulation compounds to improve insulin action in LBW subjects. Low birthweight (LBW; <2500 g) is a crude marker of a compromised intrauterine environment and has consistently been associated with insulin resistance and an increased risk of type 2 diabetes (T2D) in adult life (1). However, the exact mechanisms responsible for these changes remain unknown. The “thrifty phenotype hypothesis” (2) suggests that fetal stress or insufficient fetal nutrition induce physiological adaptations to rescue fetus survival and to prepare its organism for a hostile postnatal environment. Such adaptations, however, might simultaneously increase the risk of various metabolic and neurocognitive disturbances in adult life (2, 3). Findings from both humans and animals have indicated that LBW can cause disturbances in the limbic–hypothalamic–pituitary–adrenal (LHPA) axis regarding both diurnal and stress-related homeostasis, resulting in overall greater circulating levels of cortisol (4–6). Overexposure to corticosteroids is known to impair tissue insulin sensitivity (7); hence, LHPA axis hyperactivity might be an important factor in the development of insulin resistance and T2D in adult subjects born with LBW (5). Magnetic resonance imaging (MRI) studies of children and adults with LBW have revealed changes in limbic structures such as the hippocampus, thalamus, and prefrontal cortex (8–10), which are key limbic organs known to be involved in LHPA axis regulation (11, 12). These studies, however, did not assess LHPA axis function, and it remains unknown whether structural limbic changes are directly related to neuroendocrine disturbances within the LHPA axis. LHPA axis hyperactivity has been related to major depressive disorder, and in these patients, treatment with a selective serotonin reuptake inhibitor (SSRI) has been found to downregulate LHPA axis tonus (13, 14) and improve insulin sensitivity (15, 16). Furthermore, in LBW rats characterized by LHPA axis hyperactivity and whole body insulin resistance, we have previously shown that 4 to 5 weeks of escitalopram treatment led to complete normalization of LHPA axis regulation and whole body insulin action (17, 18). Similar studies, however, have not yet been conducted in humans with LBW. The primary aim of the present investigation was to study the limbic brain structures, LHPA axis function, and insulin sensitivity in young men with LBW vs young men with normal birthweight (NBW; part 1, the LBW phenotype study). Our secondary aim was to assess the effects of escitalopram or placebo on the LHPA axis and insulin sensitivity in LBW subjects (part 2, the LBW trial). For part 1, we recruited 40 healthy LBW men and 20 NBW men matched by body mass index (BMI), age, and physical activity. After completion of part 1, all 40 LBW subjects were enrolled in a randomized double-blinded trial with 3 months of treatment with escitalopram (20 mg/d) or placebo, followed by a complete reexamination (part 2). In both part 1 and part 2, we examined LHPA axis regulation (24-hour plasma profiles), limbic brain morphology (MRI), and insulin sensitivity (hyperinsulinemic euglycemic clamp). Materials and Methods Participants We identified and recruited 40 male LBW subjects (<2500 g) born at term through the Danish Birth Register. We prospectively recruited and matched 20 male subjects with NBW (3500 to 4500 g) at term, matched by age, BMI, and physical activity, using local and national advertising. All 60 subjects were young, healthy, and nonsmokers. They were also lean (BMI <25 kg/m2), received no medication, had no history of depression, and no first-degree relatives with T2D. All participants were white, and the parents of the subjects were of normal height (father ≥170 cm, mother ≥160 cm). Study design The present investigation consisted of two studies: the LBW phenotype study (part 1) and the LBW trial (part 2). The complete experiments (parts 1 and 2) were performed from March to September to minimize any seasonal variations. In part 1, the LBW and NBW subjects were examined at two different visits, 1 week apart. Both visits lasted for 2 days and included 2 overnight stays. All visits were initiated by an overnight stay at the clinic before the experiments to acclimatize the participants to the hospital environment. All participants received standardized meals at every visit (at 8:30 am, 12:30 pm, and 6:30 pm) and were asked to abstain from alcohol use and strenuous physical activity for 4 days before each visit. In part 2, after completion of part 1, all LBW subjects were randomized to receive double-blind treatment with placebo or escitalopram for 3 months. The dosage was 10 mg/d for week 1 and escalated to a final dose of 20 mg/d at week 2. After 3 months of intervention (placebo 101.2 ± 1.8 days, escitalopram 102.1 ± 1.9 days), all LBW participants were reexamined. All reexamination procedures and visits were performed identically to those conducted at baseline in part 1. Telephone consultations were performed after 1, 2, 4, 8, and 12 weeks to monitor for adverse events and determine drug adherence. The randomization code (generated and kept by the hospital pharmacy) was not unblinded until after the last Good Clinical Practice (GCP) audit and after the last visit of the last patient. The primary endpoints were changes in maximal insulin-stimulated glucose turnover (Rdmax), submaximal insulin-stimulated glucose turnover (Rdsubmax), and plasma-free cortisol 24-hour area under the curve (AUC). Study approval All participants submitted written informed consent before study inclusion. The study was executed in accordance with the GCP guidelines and the Declaration of Helsinki, monitored by the local GCP unit at Aarhus University Hospital, and approved by the regional ethical committees, the Danish Health and Medicines Authority, and the Danish Data Protection Agency (approval no. EudraCT 2008-004521-42; ClinicalTrials.gov identifier, NCT00971815). LHPA axis function The diurnal LHPA axis activity was evaluated using 24-hour plasma profiles of adrenocorticotropic hormone (ACTH) and cortisol, with blood samples taken using an intravenous catheter at 8:00 am, 11:00 am, 2:00 pm, 5:00 pm, 8:00 pm, 10:00 pm, 11:00 pm, 12:00 am, 2:00 am, and 5:00 am. Corticosteroid-binding globulin (CBG) was detected to calculate the plasma concentrations of free cortisol according to Coolens equation, assuming constant CBG levels (19). The total 24-hour AUC was calculated using the trapezoid method (20). Cortisol was measured by the Clinical Biochemical Department, Aarhus University Hospital, using an electrochemiluminescence immunoassay [Cobas 6000 Analyzer; Roche, Hvidovre, Denmark; coefficient of variation (intermediate precision) 3.3%]. CBG and ACTH were analyzed using commercial enzyme-linked immunosorbent assay kits EIA-3647 (DRG Instruments GmbH, Marburg, Germany) and RD19223400R (Biovendor, Brno, Czech Republic). Hyperinsulinemic euglycemic clamp Clamps were initiated by primed (20 μCi) continuous infusion (0.20 μCi/min) of D-[3-3H] glucose for 2 hours to assess the basal glucose turnover. Next, glucose turnover was assessed at both submaximal (insulin infusion rate 0.3 mU/kg/min; Actrapid; Novo Nordisk A/S, Bagsvaerd, Denmark) and maximal (1.0 mU insulin/kg/min) insulin stimulation, with a duration of 2 hours for each of the two stages. Plasma glucose was maintained during insulin stimulation at 5.0 mM using variable 20% glucose infusion. At submaximal insulin stimulation, D-[3-3H] glucose tracer was added to the infused glucose (∼100 μCi/500 mL 20% glucose) to estimate the rate of endogenous glucose production (EGP) and rate of glucose disappearance (Rd) using the Steele correction for nonsteady state (21). The basal hepatic insulin sensitivity index (HSI) was calculated as described by Miyazaki et al. (22). Plasma glucose levels were analyzed in real time using a YSI 2300 analyzer. Steady-state plasma insulin levels were assessed after clamp using a time-resolved fluoroimmunoassay (Perkin Elmer, Skovlunde, Denmark). Serum-specific activity of D-[3-3H] glucose was measured using a beta counter (TRi-Carb 2910 TR; Perkin Elmer). Body composition and MRI lipid spectroscopy The body weights and waist/hip ratios were measured by trained laboratory staff. The whole body fat content and distribution were evaluated using the same dual energy x-ray absorptiometry scanner for all subjects and visits (Hologic Discovery W; Hologic Inc., Waltham, MA). The intrahepatic lipid (IHL), intramyocellular lipid (IMCL) and extramyocellular lipid (EMCL) content of the proximal part of the right anterior tibial muscle was examined with 1H-MRI spectroscopy using a Signa Excite 1.5T twin speed scanner (GE Medical Systems, Milwaukee, WI) (23). Four muscle spectra were omitted because of technical problems (1 control subject and 3 LBW subjects). The mean full width at half maximum was 12.0 ± 0.1 Hz. The LCModel, version 6.2.1, was used to fit the individual lipid spectra and calculate the lipid/water ratio (24). Brain MRI MRI T1-weighted brain images were acquired with a three-dimensional spoiled gradient recalled scan sequence and an eight-channel, high-resolution brain coil using the following parameters: echo time, 6 ms; repetition time, 22 ms; flip angle, 30°; slice thickness, 1.3 mm; number of slices, 120 to 126; field of view, 26 to 28 cm, and acquisition matrix, 192 × 192. Two brain scans were omitted because of poor image quality (1 LBW subject and 1 control subject). The images were processed with automated segmentation using an ×86-based workstation installed with CentOS, version 6.4, 64-bit operating system, and FreeSurfer, version 5.3 (available at: http://surfer.nmr.mgh.harvard.edu) (25–27). The ventromedial prefrontal cortex (vmPFC) was defined as the sum of the medial and lateral orbitofrontal cortex (28). Longitudinal analysis was performed according to Reuter et al. (29). Physical activity Physical activity was assessed using the self-reported physical activity questionnaire (30), which was further validated in a subgroup of 11 NBW and 27 LBW men using triaxial accelerometry (ActiGraph GT3X). The total number of steps in 24 hours correlated linearly with the total self-reported physical activity questionnaire score (r = 0.54; P < 0.001). Statistical analysis Differences between the LBW and NBW subjects were tested using either an unpaired t test or the nonparametric Wilcoxon (Mann-Whitney) rank sum test according to the distribution. In part 2, group delta values (baseline vs after treatment) were analyzed using unpaired tests. The Pearson correlation test was applied to linear correlations. The STATA®/IC 11.2 statistical software package was used for statistical analysis. P values < 0.05 were considered statistically significant. Results Subject characteristics The LBW subjects were 8.8 ± 1.8 cm shorter and weighed 6.61 ± 2.15 kg less than the NBW subjects (Table 1). No differences were found in the BMI, age, or physical activity between the two groups. The LBW men exhibited a trend toward a greater waist/hip ratio (P = 0.06) and had a statistically significant greater total fat percentage (P = 0.03) and trunk fat/total fat ratio (P < 0.01). MR spectroscopy revealed statistically significantly greater IHL content level in LBW vs NBW men (P < 0.01). In contrast, no differences were found in IMCL or EMCL content. Escitalopram vs placebo in the LBW subjects did not change body composition, IHL content, muscular lipid levels, or physical activity. Table 1. Anthropometrics, Limbic Brain Morphology, and Body Composition for LBW vs NBW Men (Part 1) Variable  NBW (n = 19/20)  LBW (n = 39/40)  P Value  Anthropometric data         Birth weight, g  3775 (3600–3885)a  2350 (2150–2400)a  < 0.001   Term, wk  40 (39–41)a  38 (38–39)a  < 0.001   Age, y  25.0 ± 0.7  24.7 ± 0.5  0.76   Height, cm  185.6 ± 1.6  176.9 ± 1.0  < 0.001   Weight, kg  77.0 ± 1.9  70.4 ± 1.2  < 0.01   BMI, kg/m2  22.3 ± 0.4  22.5 ± 0.3  0.68   Physical activity,b arbitrary units  43.1 (39.5–48.1)a  42.4 (38.4–48.1)a  0.66  Body composition         Waist/hip ratio  0.85 ± 0.01  0.87 ± 0.01  0.06   Total body fat, %c  14.6 (12.9–16.5)a  17.5 (14.7–20.2)a  0.03   Trunk/total fat ratio, arbitrary unitsb  0.44 ± 0.010  0.48 ± 0.008  < 0.01   IMCL content, arbitrary units  0.88 ± 0.11  1.07 ± 0.09  0.21   EMCL content, arbitrary units  2.42 ± 0.17  2.25 ± 0.11  0.38   IHL content, arbitrary units  2.25 ± 0.14  3.00 ± 0.15  < 0.01  Limbic brain morphologyd         Left hippocampus volume, mm3  4369.9 ± 82.1  4384.2 ± 59.8  0.60   Right hippocampus volume, mm3  4338.9 ± 102.1  4444.9 ± 62.7  0.18   Left thalamus volume, mm3  8947.1 ± 255.1  8116.0 ± 111.6  < 0.01   Right thalamus volume, mm3  7923.2 ± 187.1  7671.7 ± 99.2  0.44   Left amygdala volume, mm3  1523.4 ± 34.5  1560.6 ± 26.2  0.21   Right amygdala volume, mm3  1574.1 ± 27.8  1586.3 ± 25.8  0.57   Left vmPFC volume, mm3  14,692.4 ± 362.3  14,252.6 ± 266.0  0.38   Right vmPFC volume, mm3  14,398.9 ± 314.9  14,082.9 ± 210.0  0.52   Left vmPFC surface area, mm2  4822.8 ± 125.9  4450.8 ± 84.6  < 0.01   Right vmPFC surface area, mm2  4742.2 ± 116.8  4462.3 ± 75.5  0.02   Total estimated ICV, mm3  1,712,039 ± 42,296  1,688,664 ± 21,231  0.58  Variable  NBW (n = 19/20)  LBW (n = 39/40)  P Value  Anthropometric data         Birth weight, g  3775 (3600–3885)a  2350 (2150–2400)a  < 0.001   Term, wk  40 (39–41)a  38 (38–39)a  < 0.001   Age, y  25.0 ± 0.7  24.7 ± 0.5  0.76   Height, cm  185.6 ± 1.6  176.9 ± 1.0  < 0.001   Weight, kg  77.0 ± 1.9  70.4 ± 1.2  < 0.01   BMI, kg/m2  22.3 ± 0.4  22.5 ± 0.3  0.68   Physical activity,b arbitrary units  43.1 (39.5–48.1)a  42.4 (38.4–48.1)a  0.66  Body composition         Waist/hip ratio  0.85 ± 0.01  0.87 ± 0.01  0.06   Total body fat, %c  14.6 (12.9–16.5)a  17.5 (14.7–20.2)a  0.03   Trunk/total fat ratio, arbitrary unitsb  0.44 ± 0.010  0.48 ± 0.008  < 0.01   IMCL content, arbitrary units  0.88 ± 0.11  1.07 ± 0.09  0.21   EMCL content, arbitrary units  2.42 ± 0.17  2.25 ± 0.11  0.38   IHL content, arbitrary units  2.25 ± 0.14  3.00 ± 0.15  < 0.01  Limbic brain morphologyd         Left hippocampus volume, mm3  4369.9 ± 82.1  4384.2 ± 59.8  0.60   Right hippocampus volume, mm3  4338.9 ± 102.1  4444.9 ± 62.7  0.18   Left thalamus volume, mm3  8947.1 ± 255.1  8116.0 ± 111.6  < 0.01   Right thalamus volume, mm3  7923.2 ± 187.1  7671.7 ± 99.2  0.44   Left amygdala volume, mm3  1523.4 ± 34.5  1560.6 ± 26.2  0.21   Right amygdala volume, mm3  1574.1 ± 27.8  1586.3 ± 25.8  0.57   Left vmPFC volume, mm3  14,692.4 ± 362.3  14,252.6 ± 266.0  0.38   Right vmPFC volume, mm3  14,398.9 ± 314.9  14,082.9 ± 210.0  0.52   Left vmPFC surface area, mm2  4822.8 ± 125.9  4450.8 ± 84.6  < 0.01   Right vmPFC surface area, mm2  4742.2 ± 116.8  4462.3 ± 75.5  0.02   Total estimated ICV, mm3  1,712,039 ± 42,296  1,688,664 ± 21,231  0.58  Data presented as mean ± standard error of the mean, unless otherwise indicated. Abbreviation: ICV, intracerebral volume. a Non-Gaussian variables presented as median (interquartile range). b Self-reported physical activity questionnaire score. c Estimated from dual energy x-ray absorptiometry scans. d All measures of limbic brain morphology (except for total ICV) were adjusted for total estimated ICV before statistical analyses; p values shown are from unpaired t tests. The found differences in left vmPFC surface area and left thalamus volume remained statistically significant after conservative Bonferroni correction of multiple comparison (data not shown); in contrast, the difference in right vmPFC surface area was no longer statistically significant (P = 0.016; α level = 0.005). View Large LHPA axis characterization In part 1, the 24-hour plasma hormone profiles (Fig. 1A and 1B) revealed a substantial elevation at 5:00 am for plasma levels of free cortisol (39.4 ± 4.0 vs 19.8 ± 3.4 nmol/L; P < 0.001) and ACTH (37.3 ± 4.0 vs 26.3 ± 2.4 pg/mL; P = 0.02) in LBW vs NBW subjects. Although the CBG levels were slightly decreased in LBW vs NBW men (22.9 ± 0.5 vs 24.8 ± 0.6 µg/mL; P = 0.03), the difference for the 5:00-am free cortisol level was primarily caused by a large increase in total cortisol (315.1 ± 22.1 vs 209.9 ± 23.6 nmol/L; P < 0.01). Accordingly, the 24-hour AUC of free cortisol (575.5 ± 24.1 vs 480.5 ± 32.5 nmol/L/h; P = 0.03) was 19.8% ± 8.6% greater in the LBW than in the NBW subjects (Fig. 1C), a difference that predominantly seemed driven by the 5:00 am difference. No statistically significant difference was found in the ACTH AUC over 24 hours (P = 0.75). The plasma levels of free cortisol in our data set were not related to the total or truncal fat content; hence, the differences seen remained statistically significant after adjustment for differences in whole body fat content (data not shown). Figure 1. View largeDownload slide Profiles of 24-hour plasma-free cortisol and ACTH for LBW vs NBW men (part 1). Baseline 24-hour plasma concentration curves of (A) ACTH and (B) free cortisol (NBW indicated by dotted line and white squares; LBW, by solid line and black squares). (C) Total 24-hour AUC of (left) free cortisol and (right) ACTH (NBW, white bars; LBW, black bars). Data presented as mean ± standard error of the mean (n = 20 to 40; *P < 0.05 for LBW vs NBW; **P < 0.001 for LBW vs NBW). Figure 1. View largeDownload slide Profiles of 24-hour plasma-free cortisol and ACTH for LBW vs NBW men (part 1). Baseline 24-hour plasma concentration curves of (A) ACTH and (B) free cortisol (NBW indicated by dotted line and white squares; LBW, by solid line and black squares). (C) Total 24-hour AUC of (left) free cortisol and (right) ACTH (NBW, white bars; LBW, black bars). Data presented as mean ± standard error of the mean (n = 20 to 40; *P < 0.05 for LBW vs NBW; **P < 0.001 for LBW vs NBW). In part 2, treatment of LBW with escitalopram vs placebo normalized the levels of ACTH and cortisol and led to marked reductions in 5:00 am levels of both free cortisol (−21.5 ± 5.1 vs −4.6 ± 6.3 nmol/L; P = 0.04), total cortisol (−107.9 ± 30.3 vs −12.6 ± 35.7 nmol/L; P = 0.04) and ACTH (−20.6 ± 5.7 vs 1.8 ± 3.1 pg/mL; P = 0.002; Table 2). Escitalopram vs placebo did not alter plasma CBG (P = 0.93). The 24-hour AUC of ACTH (−84.6 ± 28.5 vs 5.1 ± 20.6; P = 0.02), but not free cortisol, also decreased statistically significantly with escitalopram vs placebo (−274.5 ± 190.2 vs −111.1 ± 211.6 pg/mL; P = 0.57). Table 2. Changes After 3 Months of Treatment (Escitalopram or Placebo) in LBW Men (Part 2) Variable  Direction of Statistically Significant Differences: LBW vs NBW at Baseline  Δ Values a  Direction of Statistically Significant Changes in LBW With Escitalopram vs Placebo  P Value  LBW, Placebo (n = 20/19)  LBW, Escitalopram (n = 20/19)  Anthropometric data             BMI, kg/m2    0.075 ± 0.12  −0.065 ± 0.13    0.44   Physical activity,b arbitrary units    1.18 ± 1.62  3.03 ± 1.31    0.38  Body composition             Waist/hip ratio    −0.03 (−0.09 to −0.01)c  0.01 (−0.00 to 0.06)c    0.14   Total body fat, %  ↑  0.765 ± 0.24  0.72 ± 0.41    0.93   Trunk/total fat ratio, arbitrary units  ↑  0.002 ± 0.005  0.005 ± 0.006    0.71   IMCL content, arbitrary units    −0.29 ± 0.12  0.08 ± 0.17    0.08   EMCL content, arbitrary units    1.17 ± 0.21  1.46 ± 0.22    0.35   IHL content, arbitrary units  ↑  −0.19 ± 0.17  0.04 ± 0.09    0.24  Limbic brain morphology             Left hippocampus volume, mm3    −28.8 ± 28.2  34.1 ± 23.4    0.09   Right hippocampus volume, mm3    −59.6 ± 29.3  19.5 ± 28.7    0.06   Left thalamus volume, mm3  ↓  −58.5 ± 44.4  −23.6 ± 71.8    0.49   Right thalamus volume, mm3    21.5 ± 64.8  −3.4 ± 54.9    0.77   Left amygdala volume, mm3    −21.6 ± 16.4  14.2 ± 16.4    0.13   Right amygdala volume, mm3    16.6 ± 18.5  7.6 ± 15.8    0.71   Left vmPFC volume, mm3    −28.3 ± 108.5  80.2 ± 181.0    0.61   Right vmPFC volume, mm3    −176.4 ± 86.5  105.5 ± 188.1    0.18   Left vmPFC surface area, mm2  ↓  −21.4 ± 22.6  24.1 ± 19.2    0.14   Right vmPFC surface area, mm2  ↓  −14.2 ± 19.0  1.2 ± 36.3    0.71   Total estimated ICV, mm3    —  —    —  Liver glucose metabolism             EGPbasal, Δ in %    4.89 ± 4.47  3.55 ± 4.39    0.83   EGPsubmax, Δ in %    32.12 ± 21.10  20.00 ± 20.55    0.68   HSI, Δ in %  ↓  −11.96 ± 5.89  −2.86 ± 5.36    0.26  Peripheral glucose metabolism             Rsubmax, Δ in %  ↓  3.7 ± 5.2  23.9 ± 8.4  ↑  0.04   Rdmax, Δ in %  ↓  −3.9 ± 3.9  7.3 ± 4.9    0.08   Rdsubmax/insulin ratio, Δ in %  ↓  −3.4 ± 4.9  20.1 ± 9.1  ↑  0.03   Rdmax/insulin ratio, Δ in %  ↓  −6.5 ± 5.8  6.9 ± 6.3    0.12  LHPA axis             Free cortisol AUC-24 h, nmol/L/h  ↑  −49.1 ± 50.8  −84.6 ± 28.9    0.55   ACTH AUC 24-h, pg/mL/h    5.1 ± 20.6  −84.6 ± 28.5  ↓  0.01   5 am ACTH concentration, pg/mL  ↑  1.8 ± 3.1  −20.6 ± 5.7  ↓  0.002   5 am free cortisol concentration, nmol/L  ↑  −4.6 ± 6.3  −21.5 ± 5.1  ↓  0.04  Variable  Direction of Statistically Significant Differences: LBW vs NBW at Baseline  Δ Values a  Direction of Statistically Significant Changes in LBW With Escitalopram vs Placebo  P Value  LBW, Placebo (n = 20/19)  LBW, Escitalopram (n = 20/19)  Anthropometric data             BMI, kg/m2    0.075 ± 0.12  −0.065 ± 0.13    0.44   Physical activity,b arbitrary units    1.18 ± 1.62  3.03 ± 1.31    0.38  Body composition             Waist/hip ratio    −0.03 (−0.09 to −0.01)c  0.01 (−0.00 to 0.06)c    0.14   Total body fat, %  ↑  0.765 ± 0.24  0.72 ± 0.41    0.93   Trunk/total fat ratio, arbitrary units  ↑  0.002 ± 0.005  0.005 ± 0.006    0.71   IMCL content, arbitrary units    −0.29 ± 0.12  0.08 ± 0.17    0.08   EMCL content, arbitrary units    1.17 ± 0.21  1.46 ± 0.22    0.35   IHL content, arbitrary units  ↑  −0.19 ± 0.17  0.04 ± 0.09    0.24  Limbic brain morphology             Left hippocampus volume, mm3    −28.8 ± 28.2  34.1 ± 23.4    0.09   Right hippocampus volume, mm3    −59.6 ± 29.3  19.5 ± 28.7    0.06   Left thalamus volume, mm3  ↓  −58.5 ± 44.4  −23.6 ± 71.8    0.49   Right thalamus volume, mm3    21.5 ± 64.8  −3.4 ± 54.9    0.77   Left amygdala volume, mm3    −21.6 ± 16.4  14.2 ± 16.4    0.13   Right amygdala volume, mm3    16.6 ± 18.5  7.6 ± 15.8    0.71   Left vmPFC volume, mm3    −28.3 ± 108.5  80.2 ± 181.0    0.61   Right vmPFC volume, mm3    −176.4 ± 86.5  105.5 ± 188.1    0.18   Left vmPFC surface area, mm2  ↓  −21.4 ± 22.6  24.1 ± 19.2    0.14   Right vmPFC surface area, mm2  ↓  −14.2 ± 19.0  1.2 ± 36.3    0.71   Total estimated ICV, mm3    —  —    —  Liver glucose metabolism             EGPbasal, Δ in %    4.89 ± 4.47  3.55 ± 4.39    0.83   EGPsubmax, Δ in %    32.12 ± 21.10  20.00 ± 20.55    0.68   HSI, Δ in %  ↓  −11.96 ± 5.89  −2.86 ± 5.36    0.26  Peripheral glucose metabolism             Rsubmax, Δ in %  ↓  3.7 ± 5.2  23.9 ± 8.4  ↑  0.04   Rdmax, Δ in %  ↓  −3.9 ± 3.9  7.3 ± 4.9    0.08   Rdsubmax/insulin ratio, Δ in %  ↓  −3.4 ± 4.9  20.1 ± 9.1  ↑  0.03   Rdmax/insulin ratio, Δ in %  ↓  −6.5 ± 5.8  6.9 ± 6.3    0.12  LHPA axis             Free cortisol AUC-24 h, nmol/L/h  ↑  −49.1 ± 50.8  −84.6 ± 28.9    0.55   ACTH AUC 24-h, pg/mL/h    5.1 ± 20.6  −84.6 ± 28.5  ↓  0.01   5 am ACTH concentration, pg/mL  ↑  1.8 ± 3.1  −20.6 ± 5.7  ↓  0.002   5 am free cortisol concentration, nmol/L  ↑  −4.6 ± 6.3  −21.5 ± 5.1  ↓  0.04  Data presented as mean ± standard error of the mean, unless otherwise indicated. Abbreviation: basal, fasting conditions; ICV, intracerebral volume; max, maximal insulin stimulation (e.g., 1.0 mU/kg/min); submax, submaximal insulin stimulation (e.g., 0.3 mU/kg/min). a Absolute changes, unless otherwise indicated. b Self-reported physical activity questionnaire score. c Non-Gaussian variables presented as median (interquartile range). View Large Assessments of insulin sensitivity In part 1, the LBW and NBW men displayed comparable fasting plasma glucose levels (5.03 ± 0.08 vs 5.09 ± 0.05 mM; P = 0.47), although the LBW subjects exhibited statistically significant greater fasting insulin levels (36.4 ± 2.0 vs 27.0 ± 1.7 pmol/L; P < 0.001). The average steady-state insulin levels during the clamp were also slightly greater in the LBW group at both submaximal (130.4 ± 3.6 pM vs 105.5 ± 4.7 pM; P < 0.001) and maximal insulin stimulation (402.8 ± 9.5 pmol/L vs 335.9 ± 12.1 pmol/L; P < 0.001; Fig. 2). Because the insulin infusions were performed according to body weight; this difference might be ascribed, in part, to relatively more fat vs lean mass in the LBW vs NBW subjects. Nevertheless, despite higher steady-state insulin levels at both stage 1 (∼23.5%) and stage 2 (∼19.9%) in the LBW vs NBW groups, the glucose infusion rates at each stage were lower in the LBW vs NBW men (P < 0.05 and P < 0.01, respectively; Fig. 2), and the Rd estimates were significantly lower statistically in the LBW men (Rdsubmax, 3.51 ± 0.14 vs 4.17 ± 0.25; P = 0.01; Rdmax, 9.21 ± 0.30 vs 10.47 ± 0.31; P = 0.01; Fig. 3A). Furthermore, when the rates of glucose disposal were adjusted for higher steady-state insulin levels (Rd/insulin ratio; Fig. 3A), the differences seen in LBW vs NBW groups became even more pronounced. Together, these data strongly indicate the presence of peripheral insulin resistance in the LBW phenotype. Figure 2. View largeDownload slide Clamp rates of glucose infusion (GIR) and plasma insulin (p-insulin) concentrations for LBW vs NBW men (part 1). Primary y-axis shows GIRs to maintain euglycemia during stage 1 and 2, and secondary y-axis shows plasma insulin concentrations during basal and insulin-stimulated conditions. The last 30 minutes of each experimental period or stage were used to estimate the average steady-state plasma insulin concentration and GIR, respectively. Data presented as mean ± standard error of the mean (n = 20 to 40; *P < 0.01, LBW vs NBW; **P < 0.001, LBW vs NBW; ***P < 0.05, LBW vs NBW). Figure 2. View largeDownload slide Clamp rates of glucose infusion (GIR) and plasma insulin (p-insulin) concentrations for LBW vs NBW men (part 1). Primary y-axis shows GIRs to maintain euglycemia during stage 1 and 2, and secondary y-axis shows plasma insulin concentrations during basal and insulin-stimulated conditions. The last 30 minutes of each experimental period or stage were used to estimate the average steady-state plasma insulin concentration and GIR, respectively. Data presented as mean ± standard error of the mean (n = 20 to 40; *P < 0.01, LBW vs NBW; **P < 0.001, LBW vs NBW; ***P < 0.05, LBW vs NBW). Figure 3. View largeDownload slide Estimated rates of glucose turnover and insulin sensitivity indexes for LBW vs NBW men (part 1). (A) Basal and insulin-stimulated rates of peripheral (e.g., Rd) and (B) EGP and related estimated insulin sensitivity indexes [e.g., Rd/insulin ratios (A) and HSI (B)]. Data presented as mean ± standard error of the mean (n = 20 to 40; *P = 0.01, LBW vs NBW; **P < 0.001, LBW vs NBW; ***P < 0.0001, LBW vs NBW; ****P < 0.03, LBW vs NBW). Basal, fasting conditions; max, maximal insulin stimulation (e.g., 1.0 mU/kg/min); submax, submaximal insulin stimulation (e.g., 0.3 mU/kg/min). Figure 3. View largeDownload slide Estimated rates of glucose turnover and insulin sensitivity indexes for LBW vs NBW men (part 1). (A) Basal and insulin-stimulated rates of peripheral (e.g., Rd) and (B) EGP and related estimated insulin sensitivity indexes [e.g., Rd/insulin ratios (A) and HSI (B)]. Data presented as mean ± standard error of the mean (n = 20 to 40; *P = 0.01, LBW vs NBW; **P < 0.001, LBW vs NBW; ***P < 0.0001, LBW vs NBW; ****P < 0.03, LBW vs NBW). Basal, fasting conditions; max, maximal insulin stimulation (e.g., 1.0 mU/kg/min); submax, submaximal insulin stimulation (e.g., 0.3 mU/kg/min). Regarding the rates of EGP, we were unable to detect any differences between the LBW vs NBW groups, neither in the basal nor in the submaximally insulin-stimulated state (Fig. 3B), and no difference was found in the insulin-mediated relative suppression of EGP (data not shown). However, the HSI was significantly lower statistically in the LBW vs NBW men [5.60 ± 0.37 vs 7.06 ± 0.60 (mg/kg/min) × pmol/L × 10−4; P = 0.03], suggesting that hepatic insulin resistance might have been overlooked in our clamp experiment owing to the higher insulin levels. In addition to these findings, the 20 LBW subjects with 5:00 am free cortisol greater than the median (60.2 ± 3.6 nmol/L) had a statistically significant lower Rd at submaximal insulin stimulation (3.16 ± 0.14 vs 3.85 ± 0.21 mg/kg/min; P < 0.01) and a lower basal HSI (4.53 ± 0.36 vs 6.67 ± 0.55 [10,000/(mg/kg/min) × pM]; P = 0.002) compared with the 20 LBW subjects with a 5:00 am free cortisol less than the median (19.5 ± 2.8 nmol/L), suggesting even more pronounced insulin resistance in the LBW subjects with the highest free cortisol levels. No differences were found in age, BMI, truncal fat percentage or physical activity between these two subgroups of LBW individuals (data not shown). In the clamps performed in part 2 on LBW subjects exposed to either escitalopram or placebo, respectively, the insulin levels were the same (data not shown). Treatment with escitalopram vs placebo significantly increased (with statistical significance) the Rdsubmax and Rdsubmax/insulin ratio in LBW subjects at submaximal insulin stimulation by ∼24% and ∼20% (23.9% ± 8.4% vs 3.7% ± 5.2%, P = 0.04; 20.1% ± 9.1% vs −3.4% ± 4.9%, P = 0.03; Table 2). Although a trend was detected, no statistically significant effect of escitalopram vs placebo on the Rdmax or Rdmax/insulin ratio was obtained during maximal insulin stimulation (7.3% ± 4.9% vs −3.9% ± 3.9%, P = 0.08; 6.9% ± 6.3% vs −6.5% ± 5.8%, P = 0.12). Limbic brain morphology LBW vs NBW subjects (part 1) were characterized by marked reductions in the volume of the left thalamus (8.2% ± 2.9%; P < 0.01) and surface areas of both left (6.8% ± 2.3%; P < 0.01) and right (4.9% ± 2.0%; P = 0.02) vmPFC (Table 1). These specific limbic measures further correlated inversely with the plasma levels of 5:00 am free cortisol (Fig. 4A–4C; left thalamus volume, r = −0.37, P = 0.02; right vmPFC surface area, r = −0.39, P = 0.02; left vmPFC surface area, r = −0.41; P = 0.01). The correlations with 24-hour free cortisol AUC were much weaker and statistically nonsignificant (data not shown). However, no differences were found between the LBW and NBW subjects in the volumes of the right thalamus, amygdalae, or hippocampi (Table 1). Figure 4. View largeDownload slide Correlations of 5:00 am free cortisol to key limbic brain measures for LBW men (part 1). Pearson linear correlations between 5:00 am free cortisol and key limbic measures in LBW subjects. Scatter plots for 5:00 am free cortisol vs surface areas of (A) left and (B) right vmPFC. (C) Scatter plot for 5:00 am free cortisol vs left thalamus volume. The related correlation coefficients (r) and p values are indicated in the upper right corner of each panel; n = 39 for each plot. Figure 4. View largeDownload slide Correlations of 5:00 am free cortisol to key limbic brain measures for LBW men (part 1). Pearson linear correlations between 5:00 am free cortisol and key limbic measures in LBW subjects. Scatter plots for 5:00 am free cortisol vs surface areas of (A) left and (B) right vmPFC. (C) Scatter plot for 5:00 am free cortisol vs left thalamus volume. The related correlation coefficients (r) and p values are indicated in the upper right corner of each panel; n = 39 for each plot. In the LBW subjects exposed for 3 months to escitalopram vs placebo (part 2), longitudinal analyses could not reveal any statistically significant changes in the cortical volumes, subcortical volumes, or surface areas (Table 2). Discussion In part 1 of our investigation (LBW phenotype study), we found that young LBW men compared with the matched NBW controls exhibited an ∼20% increment in 24-hour free cortisol AUC, primarily driven by a prominent increase in the early morning/late night (5:00 am) plasma-free cortisol with an equivalent increase in plasma ACTH levels. In humans, the limbic brain structures such as the hippocampus, amygdala, and vmPFC and the dorsal, anteroventral, and paraventricular nuclei of the thalamus are known to be involved in the intrinsic regulation of both circadian rhythms and the reactivity of the LHPA axis (11, 12). Moreover, the thalamus seems to play an important role in the adaptive stress response (31), at same time, modulating both diurnal and stress-related hypothalamic release of corticotrophin-releasing hormone (32). We detected reductions in the volume of the left thalamus and the surface areas of the left and right vmPFC. These specific limbic measures further correlated inversely with elevated 5:00 am free-plasma cortisol levels but not with the 24-hour cortisol AUC. Hence, it appears that the limbic changes were primarily linked to late nighttime hyperactivity of the LHPA axis rather than 24-hour LHPA axis hyperactivity. Furthermore, as glucocorticoids stimulate lipolysis, our present finding of a specific increase in late night cortisol could be in line with a report by Reuter et al. (29) demonstrating diurnal differences in the rates of lipolysis, with greater rates during the night than during the day in young LBW men. Moreover, although LHPA axis disturbances previously have been reported in LBW subjects (4, 5, 18), to the best of our knowledge, the present study is the first to provide evidence that LHPA axis hyperactivity in LBW individual might be attributable to structural anomalies within key limbic organs such as the thalamus and vmPFC. Glucocorticoids are known to cause insulin resistance in skeletal muscle tissue (7), and the present LBW subjects displayed peripheral insulin resistance with lower rates of glucose disappearance (Rd) at both submaximal and maximal insulin stimulation. Moreover, insulin resistance was substantially worse in those LBW men within the upper half of the range of 5:00 am free cortisol levels. Hence, based on our findings from part 1, it would be tempting for us to hypothesize that a causal relationship might exist between LHPA axis hyperactivity and insulin resistance in LBW individuals. Although lipid accumulation in skeletal muscle tissues has been associated with insulin resistance (33), the insulin-resistant LBW vs NBW men did not display any increases in either IMCL or EMCL content. A similar disassociation between muscle fat and muscle insulin action has been reported by Dufour and Petersen (34) in LBW humans. A possible explanation for this might be that glucocorticoids seem to both stimulate intramyocellular lipolysis and inhibit lipogenesis (35), thereby potentially minimizing lipid accumulation in muscle tissue despite increased whole body fat content. Both basal and insulin-stimulated EGP rates and the relative degree of insulin-mediated EGP suppression were comparable between LBW and NBW subjects. Nevertheless, it is likely that the hepatic insulin sensitivity was overestimated in the LBW men owing to the greater insulin levels during the clamp. Hence, we decided, in addition, to calculate the basal HSI (22), a validated alternative measure of hepatic insulin action. The basal HSI was substantially lower in the LBW than in the NBW men. In addition, we found increased levels of IHL content in the livers of the LBW compared with the NBW subjects, a finding that consistently has been linked to hepatic insulin resistance (36). Hence, overall, the presented data could still be in support of previous findings of liver insulin resistance in humans with a LBW (37). In part 2 of our investigation (the LBW trial), the LBW subjects were treated with escitalopram or placebo for 3 months in a randomized and double-blind manner. The intervention with escitalopram was capable of normalizing the plasma levels of 5:00 am free cortisol and ACTH and, at the same time, vastly improving peripheral insulin sensitivity by ∼24% as measured at submaximal insulin stimulation. However, insulin responsiveness at maximal insulin stimulation was only slightly improved without reaching statistical significance. Furthermore, these improvements in insulin action and LHPA axis homeostasis were not related to changes in body composition or physical activity. Similar beneficial effects with SSRI compounds on the LHPA axis and insulin action have been reported by us in a previous study performed on LBW rats (18) and by others of patients with major depressive disorder (15). In addition, Breum et al. (38) demonstrated that SSRI treatment of obese patients with T2D improved glycemic control independently of weight loss. The investigators speculated that the insulin-sensitizing effect might have resulted from upregulation of glycogen synthase activity in skeletal muscle. This is an interesting view, because it has been shown that cortisol exposure of skeletal muscle tissue leads to insulin resistance, accompanied by a marked reduction in glycogen synthase activity (39). Thus, it can be speculated further that the beneficial effects of SSRI treatment in the study by Breum et al. (38) might have resulted from reduced cortisol exposure, although this was not measured. Although a direct beneficial effect of escitalopram on muscle tissue insulin action could not be ruled out, data are available that speak against this notion. First, a previous study reported that serotonin receptor activation in human skeletal muscle had no effect on insulin-mediated glucose uptake (40). Second, in pilot studies, we have previously seen that the effects of escitalopram regarding insulin action were highly phenotype dependent, with a negative effect on glucose metabolism (and the LHPA axis) in NBW rats, despite positive effects in LBW rats (18). Together, these data speak against the idea that escitalopram exerts a direct insulin-sensitizing effect on the peripheral tissues via local serotonin receptor activation. Escitalopram exposure did not completely restore insulin action in LBW subjects. However, escitalopram exposure also failed to improve the adverse body composition and fatty liver seen in LBW subjects. These remaining negative metabolic defects seen after normalization of cortisol levels might contribute to the residual insulin resistance. Nevertheless, when the positive metabolic and neuroendocrine effects of escitalopram seen in the LBW trial (part 2) are considered with the findings of insulin resistance and increased free cortisol levels in the LBW phenotype study (part 1), our results indicate that escitalopram improved insulin action in LBW subjects via a reduction in cortisol. Hence, the findings from part 2 provide further support for our hypothesis that hypercortisolemia might play an etiological role in the development of insulin resistance in LBW subjects. In conclusion, we have presented evidence that young men born with LBW exhibit morphological changes within key components of the limbic system that are linked to increased late night or early morning secretion of ACTH and cortisol. Furthermore, the presented data suggest that these neuroendocrine changes might, at least in part, be causally associated to the development of peripheral insulin resistance. Finally, the present investigation has shown that an SSRI compound can be used to reverse LHPA axis hyperactivity and ameliorate peripheral insulin resistance in LBW humans. Hence, compounds ameliorating insulin resistance by targeting defects in the LHPA axis could be clinically useful for reducing the risk of T2D in individuals born with a LBW. Abbreviations: ACTH adrenocorticotropic hormone AUC area under the curve BMI body mass index CBG corticosteroid-binding globulin EGP endogenous glucose production GCP Good Clinical Practice HSI hepatic insulin sensitivity index LBW low birthweight LHPA limbic–hypothalamic–pituitary–adrenal MRI magnetic resonance imaging NBW normal birthweight Rd rate of glucose disappearance Rdmax maximal insulin-stimulated glucose turnover Rdsubmax submaximal insulin-stimulated glucose turnover SSRI selective serotonin reuptake inhibitor T2D type 2 diabetes vmPFC ventromedial prefrontal cortex. Acknowledgments We thank the study participants and the laboratory technicians Eva Schriver, Karen Mathiassen, Lone Kvist, Dorthe Wulf, and Joan Hansen for their contributions. Special thanks to Annette Mengel for help with the hyperinsulinemic euglycemic clamps. Thanks also to Professor Kitt Falk Petersen, Department of Internal Medicine, Yale University School of Medicine, New Haven, Connecticut, for inspiring scientific discussions. Financial Support: The present study was funded by an independent investigator grant from H. Lundbeck A/S, a PhD grant from Aarhus University, the Augustinusfonden and Aase and Ejnar Danielsen Fond. H. Lundbeck A/S supplied the trial medication. Clinical Trial Information: ClinicalTrials.gov identifier, NCT00971815 (registered 4 September 2009). Disclosure Summary: The study was investigator initiated, and, apart from the independent investigator grant from H. Lundbeck A/S, the authors have nothing to disclose. References 1. Harder T, Rodekamp E, Schellong K, Dudenhausen JW, Plagemann A. Birth weight and subsequent risk of type 2 diabetes: a meta-analysis. Am J Epidemiol . 2007; 165( 8): 849– 857. Google Scholar CrossRef Search ADS PubMed  2. Hales CN, Barker DJ. The thrifty phenotype hypothesis. Br Med Bull . 2001; 60( 1): 5– 20. Google Scholar CrossRef Search ADS PubMed  3. Meher S, Hernandez-Andrade E, Basheer SN, Lees C. Impact of cerebral redistribution on neurodevelopmental outcome in small-for-gestational-age or growth-restricted babies: a systematic review. Ultrasound Obstet Gynecol . 2015; 46( 4): 398– 404. Google Scholar CrossRef Search ADS PubMed  4. Phillips DI, Walker BR, Reynolds RM, Flanagan DE, Wood PJ, Osmond C, Barker DJ, Whorwood CB. Low birth weight predicts elevated plasma cortisol concentrations in adults from 3 populations. Hypertension . 2000; 35( 6): 1301– 1306. Google Scholar CrossRef Search ADS PubMed  5. Phillips DI, Barker DJ, Fall CH, Seckl JR, Whorwood CB, Wood PJ, Walker BR. Elevated plasma cortisol concentrations: a link between low birth weight and the insulin resistance syndrome? J Clin Endocrinol Metab . 1998; 83( 3): 757– 760. Google Scholar PubMed  6. Kajantie E, Feldt K, Räikkönen K, Phillips DI, Osmond C, Heinonen K, Pesonen AK, Andersson S, Barker DJ, Eriksson JG. Body size at birth predicts hypothalamic-pituitary-adrenal axis response to psychosocial stress at age 60 to 70 years. J Clin Endocrinol Metab . 2007; 92( 11): 4094– 4100. Google Scholar CrossRef Search ADS PubMed  7. Weinstein SP, Paquin T, Pritsker A, Haber RS. Glucocorticoid-induced insulin resistance: dexamethasone inhibits the activation of glucose transport in rat skeletal muscle by both insulin- and non-insulin-related stimuli. Diabetes . 1995; 44( 4): 441– 445. Google Scholar CrossRef Search ADS PubMed  8. Lowe J, Duvall SW, MacLean PC, Caprihan A, Ohls R, Qualls C, Phillips J. Comparison of structural magnetic resonance imaging and development in toddlers born very low birth weight and full-term. J Child Neurol . 2011; 26( 5): 586– 592. Google Scholar CrossRef Search ADS PubMed  9. Martinussen M, Flanders DW, Fischl B, Busa E, Løhaugen GC, Skranes J, Vangberg TR, Brubakk AM, Haraldseth O, Dale AM. Segmental brain volumes and cognitive and perceptual correlates in 15-year-old adolescents with low birth weight. J Pediatr . 2009; 155( 6): 848– 853.e1. Google Scholar CrossRef Search ADS PubMed  10. Skranes J, Løhaugen GC, Martinussen M, Håberg A, Brubakk AM, Dale AM. Cortical surface area and IQ in very-low-birth-weight (VLBW) young adults. Cortex . 2013; 49( 8): 2264– 2271. Google Scholar CrossRef Search ADS PubMed  11. Ahima R, Krozowski Z, Harlan R. Type I corticosteroid receptor-like immunoreactivity in the rat CNS: distribution and regulation by corticosteroids. J Comp Neurol . 1991; 313( 3): 522– 538. Google Scholar CrossRef Search ADS PubMed  12. Morimoto M, Morita N, Ozawa H, Yokoyama K, Kawata M. Distribution of glucocorticoid receptor immunoreactivity and mRNA in the rat brain: an immunohistochemical and in situ hybridization study. Neurosci Res . 1996; 26( 3): 235– 269. Google Scholar CrossRef Search ADS PubMed  13. Piwowarska J, Chimiak A, Matsumoto H, Dziklińska A, Radziwoń-Zaleska M, Szelenberger W, Pachecka J. Serum cortisol concentration in patients with major depression after treatment with fluoxetine. Psychiatry Res . 2012; 198( 3): 407– 411. Google Scholar CrossRef Search ADS PubMed  14. Nikisch G, Mathé AA, Czernik A, Thiele J, Bohner J, Eap CB, Agren H, Baumann P. Long-term citalopram administration reduces responsiveness of HPA axis in patients with major depression: relationship with S-citalopram concentrations in plasma and cerebrospinal fluid (CSF) and clinical response. Psychopharmacology (Berl) . 2005; 181( 4): 751– 760. Google Scholar CrossRef Search ADS PubMed  15. Weber-Hamann B, Gilles M, Lederbogen F, Heuser I, Deuschle M. Improved insulin sensitivity in 80 nondiabetic patients with MDD after clinical remission in a double-blind, randomized trial of amitriptyline and paroxetine. J Clin Psychiatry . 2006; 67( 12): 1856– 1861. Google Scholar CrossRef Search ADS PubMed  16. Lustman PJ, Freedland KE, Griffith LS, Clouse RE. Fluoxetine for depression in diabetes: a randomized double-blind placebo-controlled trial. Diabetes Care . 2000; 23( 5): 618– 623. Google Scholar CrossRef Search ADS PubMed  17. Buhl ES, Neschen S, Yonemitsu S, Rossbacher J, Zhang D, Morino K, Flyvbjerg A, Perret P, Samuel V, Kim J, Cline GW, Petersen KF. Increased hypothalamic-pituitary-adrenal axis activity and hepatic insulin resistance in low-birth-weight rats. Am J Physiol Endocrinol Metab . 2007; 293( 5): E1451– E1458. Google Scholar CrossRef Search ADS PubMed  18. Buhl ES, Jensen TK, Jessen N, Elfving B, Buhl CS, Kristiansen SB, Pold R, Solskov L, Schmitz O, Wegener G, Lund S, Petersen KF. Treatment with an SSRI antidepressant restores hippocampo-hypothalamic corticosteroid feedback and reverses insulin resistance in low-birth-weight rats. Am J Physiol Endocrinol Metab . 2010; 298( 5): E920– E929. Google Scholar CrossRef Search ADS PubMed  19. Coolens JL, Van Baelen H, Heyns W. Clinical use of unbound plasma cortisol as calculated from total cortisol and corticosteroid-binding globulin. J Steroid Biochem . 1987; 26( 2): 197– 202. Google Scholar CrossRef Search ADS PubMed  20. Fall CH, Dennison E, Cooper C, Pringle J, Kellingray SD, Hindmarsh P. Does birth weight predict adult serum cortisol concentrations? Twenty-four-hour profiles in the United kingdom 1920-1930 Hertfordshire Birth Cohort. J Clin Endocrinol Metab . 2002; 87( 5): 2001– 2007. Google Scholar PubMed  21. Steele R. Influences of glucose loading and of injected insulin on hepatic glucose output. Ann N Y Acad Sci . 1959; 82( 2): 420– 430. Google Scholar CrossRef Search ADS PubMed  22. Miyazaki Y, Glass L, Triplitt C, Wajcberg E, Mandarino LJ, DeFronzo RA. Abdominal fat distribution and peripheral and hepatic insulin resistance in type 2 diabetes mellitus. Am J Physiol Endocrinol Metab . 2002; 283( 6): E1135– E1143. Google Scholar CrossRef Search ADS PubMed  23. Wamberg L, Kampmann U, Stødkilde-Jørgensen H, Rejnmark L, Pedersen SB, Richelsen B. Effects of vitamin D supplementation on body fat accumulation, inflammation, and metabolic risk factors in obese adults with low vitamin D levels—results from a randomized trial. Eur J Intern Med . 2013; 24( 7): 644– 649. Google Scholar CrossRef Search ADS PubMed  24. Provencher SW. Estimation of metabolite concentrations from localized in vivo proton NMR spectra. Magn Reson Med . 1993; 30( 6): 672– 679. Google Scholar CrossRef Search ADS PubMed  25. Morey RA, Petty CM, Xu Y, Hayes JP, Wagner HR II, Lewis DV, LaBar KS, Styner M, McCarthy G. A comparison of automated segmentation and manual tracing for quantifying hippocampal and amygdala volumes. Neuroimage . 2009; 45( 3): 855– 866. Google Scholar CrossRef Search ADS PubMed  26. Keller SS, Gerdes JS, Mohammadi S, Kellinghaus C, Kugel H, Deppe K, Ringelstein EB, Evers S, Schwindt W, Deppe M. Volume estimation of the thalamus using FreeSurfer and stereology: consistency between methods. Neuroinformatics . 2012; 10( 4): 341– 350. Google Scholar CrossRef Search ADS PubMed  27. Dale AM, Fischl B, Sereno MI. Cortical surface-based analysis. I. Segmentation and surface reconstruction. Neuroimage . 1999; 9( 2): 179– 194. Google Scholar CrossRef Search ADS PubMed  28. Boes AD, Bechara A, Tranel D, Anderson SW, Richman L, Nopoulos P. Right ventromedial prefrontal cortex: a neuroanatomical correlate of impulse control in boys. Soc Cogn Affect Neurosci . 2009; 4( 1): 1– 9. Google Scholar CrossRef Search ADS PubMed  29. Reuter M, Schmansky NJ, Rosas HD, Fischl B. Within-subject template estimation for unbiased longitudinal image analysis. Neuroimage . 2012; 61( 4): 1402– 1418. Google Scholar CrossRef Search ADS PubMed  30. Aadahl M, Jørgensen T. Validation of a new self-report instrument for measuring physical activity. Med Sci Sports Exerc . 2003; 35( 7): 1196– 1202. Google Scholar CrossRef Search ADS PubMed  31. Jaferi A, Bhatnagar S. Corticosterone can act at the posterior paraventricular thalamus to inhibit hypothalamic-pituitary-adrenal activity in animals that habituate to repeated stress. Endocrinology . 2006; 147( 10): 4917– 4930. Google Scholar CrossRef Search ADS PubMed  32. Hsu DT, Bakshi VP, Roseboom PH, Kalin NH. Diurnal changes in corticotropin-releasing hormone messenger RNA in the rat thalamus. Neurosci Lett . 2003; 338( 1): 33– 36. Google Scholar CrossRef Search ADS PubMed  33. Pan DA, Lillioja S, Kriketos AD, Milner MR, Baur LA, Bogardus C, Jenkins AB, Storlien LH. Skeletal muscle triglyceride levels are inversely related to insulin action. Diabetes . 1997; 46( 6): 983– 988. Google Scholar CrossRef Search ADS PubMed  34. Dufour S, Petersen KF. Disassociation of liver and muscle insulin resistance from ectopic lipid accumulation in low-birth-weight individuals. J Clin Endocrinol Metab . 2011; 96( 12): 3873– 3880. Google Scholar CrossRef Search ADS PubMed  35. Morgan SA, Gathercole LL, Simonet C, Hassan-Smith ZK, Bujalska I, Guest P, Abrahams L, Smith DM, Stewart PM, Lavery GG, Tomlinson JW. Regulation of lipid metabolism by glucocorticoids and 11β-HSD1 in skeletal muscle. Endocrinology . 2013; 154( 7): 2374– 2384. Google Scholar CrossRef Search ADS PubMed  36. Petersen MC, Shulman GI. Roles of diacylglycerols and ceramides in hepatic insulin resistance. Trends Pharmacol Sci.  2017; 38: 649– 665. Google Scholar CrossRef Search ADS PubMed  37. Brøns C, Jensen CB, Storgaard H, Alibegovic A, Jacobsen S, Nilsson E, Astrup A, Quistorff B, Vaag A. Mitochondrial function in skeletal muscle is normal and unrelated to insulin action in young men born with low birth weight. J Clin Endocrinol Metab . 2008; 93( 10): 3885– 3892. Google Scholar CrossRef Search ADS PubMed  38. Breum L, Bjerre U, Bak JF, Jacobsen S, Astrup A. Long-term effects of fluoxetine on glycemic control in obese patients with non-insulin-dependent diabetes mellitus or glucose intolerance: influence on muscle glycogen synthase and insulin receptor kinase activity. Metabolism . 1995; 44( 12): 1570– 1576. Google Scholar CrossRef Search ADS PubMed  39. Holmäng A, Björntorp P. The effects of cortisol on insulin sensitivity in muscle. Acta Physiol Scand . 1992; 144( 4): 425– 431. Google Scholar CrossRef Search ADS PubMed  40. Hajduch E, Rencurel F, Balendran A, Batty IH, Downes CP, Hundal HS. Serotonin (5-Hydroxytryptamine), a novel regulator of glucose transport in rat skeletal muscle. J Biol Chem . 1999; 274( 19): 13563– 13568. Google Scholar CrossRef Search ADS PubMed  Copyright © 2018 Endocrine Society

Journal

Journal of Clinical Endocrinology and MetabolismOxford University Press

Published: Jan 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Unlimited reading

Read as many articles as you need. Full articles with original layout, charts and figures. Read online, from anywhere.

Stay up to date

Keep up with your field with Personalized Recommendations and Follow Journals to get automatic updates.

Organize your research

It’s easy to organize your research with our built-in tools.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve Freelancer

DeepDyve Pro

Price
FREE
$49/month

$360/year
Save searches from
Google Scholar,
PubMed
Create lists to
organize your research
Export lists, citations
Read DeepDyve articles
Abstract access only
Unlimited access to over
18 million full-text articles
Print
20 pages/month
PDF Discount
20% off