Disassociation of Liver and Muscle Insulin Resistance from Ectopic Lipid Accumulation in Low-Birth-Weight Individuals

Abstract

Context:

Low birth weight (LBW) is a marker of fetal stress and is associated with an increased prevalence of type 2 diabetes (T2D). Insulin resistance plays a prominent role in the development of T2D; however, the pathogenesis of T2D in LBW is controversial.

Objective:

The objective of the study was to assess whole-body and tissue-specific insulin sensitivity and intramyocellular lipid (IMCL) and hepatic lipid content in LBW and matched control subjects.

Design:

These were prospective and pair-matched studies.

Setting:

The study was conducted at Yale University Center for Clinical Investigation.

Participants:

Young, lean, nonsmoking, sedentary LBW (n = 45) and matched control subjects participated in the study.

Intervention:

Interventions included an oral glucose tolerance test and hyperinsulinemic-euglycemic clamps and ¹H magnetic resonance spectroscopy.

Main Outcomes Measures:

The main outcomes measures included insulin sensitivity index; whole-body and tissue-specific insulin sensitivity; liver lipid and IMCL contents; and fasting concentrations of cortisol, GH, and IGF-I as markers of the hypothalamic-pituitary-adrenal and IGFI/GH axes.

Results:

The LBW subjects were insulin resistant as reflected by a 20% reduction in insulin sensitivity index as compared with the controls (P = 0.0017), which could be attributed to both liver and muscle insulin resistance. There were no differences in IMCL or hepatic triglyceride content between LBW and control groups. In the LBW group, fasting plasma concentrations of cortisol (P = 0.01) and GH (P = 0.01) were increased, and IGF-I concentrations reduced (P < 0.05) a pattern, which may suggest potential dysregulation of the hypothalamic-pituitary-adrenal and IGF-I/GH axes.

Conclusion:

These results support the hypothesis that fetal stress and LBW lead to liver and muscle insulin resistance and show that this is independent of lipid deposition in these organs.

Type 2 diabetes (T2D) is highly prevalent in subjects born with low birth weight (LBW), currently defined as less than 5 lb 8 oz (1–4); however, it is unclear whether LBW is associated with insulin resistance. Most studies addressing insulin sensitivity in LBW have yielded conflicting results, and the studies suggesting insulin resistance in LBW individuals have not controlled for confounding factors such as smoking, body weight, and physical activity (5–10). In addition, no studies have determined whether LBW is associated with ectopic lipid deposition in liver and skeletal muscle, which have been shown to be strongly associated with insulin resistance in obesity, first-degree relatives of T2D individuals, T2D patients, and the elderly (11–23).

LBW is a marker of an adverse fetal environment and fetal stress, and one hypothesis put forward by Barker (4) suggests that harmful events during the fetal period can induce life-long changes in different organs predisposing to development of disease, possibly through resetting of a diverse array of hormonal systems. Of particular interest is the stress response, especially the regulation of the hypothalamic-pituitary-adrenal (HPA) axis, which has potent effects on metabolism and vasculature (24). In a rat model of LBW and intrauterine stress, we recently demonstrated insulin resistance and accompanying dysregulation of the HPA axis with chronically excessive adrenal glucocorticoid secretion and increased stress responses (25).

A few studies suggest that in humans there may be a relationship between birth weight and a biological response to stress (26, 27), and a large study by Phillips and colleagues (24, 26) suggested an inverse relation between birth weight and fasting plasma cortisol levels; however, no definitive conclusions about birth weight, stress response, and a potential link to insulin resistance and T2D have been reached.

In the present study, we examined whole-body as well as tissue-specific insulin responsiveness of liver, muscle, and fat cells and examined the potential relation to ectopic lipid accumulation in muscle and liver using ¹H magnetic resonance spectroscopy (MRS). In addition, we examined the relation between insulin responsiveness in these organs and markers of the GH/IGF-I axis and the HPA axis.

Materials and Methods

All subjects were recruited prospectively via local advertisement for the oral glucose tolerance test (OGTT) and ¹H MRS studies. From these subjects in the LBW group, 15 subjects who agreed to participate, were enrolled in the clamp studies, and 15 control subjects who agreed to participate and matched to the LBW group for gender, body mass index (BMI), and fat mass were selected. Inclusion criteria were age 18–40 yr, BMI 19.0–26.9 kg/m², nonsmoking, taking no medications (except for birth control), and a sedentary lifestyle as defined by an activity questionnaire (28) combined with a sedentary job. Physical activity was monitored for 3 consecutive days by pedometer (GOWalking; Sportline Inc., Hazleton, PA) (12, 13). Subjects were divided into two qualifying groups: LBW: 0.9–2.5 kg (average 2.0 ± 0.1 kg) and control: 2.8–4.5 kg (average 3.4 ± 0.1 kg) born at term (wk 39–41). The LBW subjects provided birth certificates to document birth weight and term.

Written consent was obtained from each subject after the purpose, nature, and potential complications of the studies had been explained. The protocol was approved by the Human Investigation Committee at Yale University.

Subjects were instructed to eat a regular, weight-maintenance diet containing at least 150 g of carbohydrate per day and not to perform any exercise other than normal walking for the 3 d before the OGTT and the euglycemic-hyperinsulinemic clamp studies. Women were studied in the follicular phase of the menstrual cycle (between d 0 and 12) to eliminate changes in metabolism due to hormonal changes during the menstrual cycle (29).

For the OGTT, subjects were admitted in the morning after an overnight fast at the Yale Center for Clinical Investigation Hospital Research Unit as previously described (12, 13, 15). Twenty minutes after insertion of an antecubital intravenous line, fasting blood samples were collected for determination of fasting plasma substrate and hormone concentrations and blood samples were collected at 10, 20, 30, 60, 90, 120, 150, and 180 min for the determination of plasma glucose, C-peptide, and insulin concentrations. Insulin sensitivity was assessed by the insulin sensitivity index (ISI) (10⁻⁴ dl/min per milliunits per milliliter) using the oral glucose minimal model (15). This ISI has been shown to reflect insulin sensitivity in large populations and measures overall effects of insulin to stimulate glucose disposal and inhibit glucose production (15).

On a separate day, control and LBW subjects underwent localized ¹H MRS on a 4T whole-body magnet interfaced to a Bruker AVANCE Spectrometer (Bruker Instruments, Billerica, MA). Intramyocellular lipid (IMCL) content of the calf muscles was measured by using a STEAM sequence, in conjunction with a ¹H-quadrature probe with twin 13-cm coils as described (15). Localized ¹H MRS spectra of the liver were obtained by using a STEAM sequence with respiratory gating, outer volume suppression as described (15) and T2 correction.

For the insulin clamp studies, the subjects were admitted to the Yale Center for Clinical Investigation Hospital Research Unit the evening before the study and fasted overnight. A primed-continuous infusion of [6,6-²H₂] glucose and a constant infusion of [²H₅] glycerol were initiated in the morning and continued throughout the study as previously described (13). At the end of the 3-h baseline period and during the 3-h insulin clamp, blood samples were collected and analyzed for glucose, insulin, C-peptide concentrations, and plasma enrichments of [6,6-²H₂] glucose and [²H₅] glycerol. Tissue-specific insulin sensitivity was assessed with euglycemic-hyperinsulinemic clamps of 20 mU/(m²·min) specifically for muscle (12, 13) and of 10 mU/(m²·min) specifically for liver. These insulin levels allow for detection of subtle differences in muscle, adipocyte, and liver insulin sensitivity compared with using higher insulin levels (12, 13, 23).

Lean body and fat masses of the subgroups in the insulin clamp studies were measured by dual-energy x-ray absorptiometry scanning and bioelectrical impedance (17).

Rates of whole-body energy expenditure, glucose, and fat oxidation in the subjects who underwent the insulin clamp studies were assessed by indirect calorimetry (Deltratrack metabolic monitor; Sensormedics, Anaheim, CA) as previously described (12, 13, 30, 31).

Plasma glucose concentrations were measured as previously described (12, 13). Plasma concentrations of insulin, C-peptide, adiponectin, total IGF-I binding protein (BP)-3, cortisol, and GH were measured using double-antibody RIA kits (Linco, St. Louis, MO). Plasma concentrations of IL-6 and TNF-α were measured by Quantine high-sensitivity kits (R&D Systems, Inc., Minneapolis, MN). Gas chromatography mass spectrometer analyses of plasma enrichments of [6,6-²H₂]-glucose and [²H₅]-glycerol were performed as previously described (13).

Calculations

The ISI was calculated from the plasma glucose and insulin concentrations before and during the OGTT as previously described (15).

Basal and insulin stimulated rates of glucose metabolism and production were calculated as the ratio of tracer to tracee (32, 33) using the following formulas. The formula for the rates of basal glucose production was: [basal (6,6-²H₂) glucose infusion rate (milligrams per minute)/BW (kilograms)] × ([infusate enrichment/average plasma enrichment] − 1).

The formula for the rates of clamped glucose production during the insulin infusion was: [mean glucose infusion rate (milligrams per minute)/BW (kilograms) during the clamp procedure] × [(infusate enrichment/average plasma enrichment) − 1].

Rates of glucose infusion were calculated in 20-min blocks between 80 and 120 min during the 20 mU/(m²·min) insulin clamp and between 120 and 180 min of the 10 mU/(m²·min) insulin clamp as previously described (12–14). Data were corrected for glucose space and averaged.

Rates of nonoxidative glucose metabolism were calculated as the difference between rates of whole-body glucose metabolism and rates of glucose oxidation.

Basal and insulin suppression of rates of glycerol release were calculated as the ratio of tracer to tracee as previously described and expressed per kilogram fat mass (see Fig. 4) (12).

Fig. 4.

Rates of glycerol turnover in the basal period and the hyperinsulinemic-euglycemic [20 mU/(kilograms of fat mass per minute)] clamp.

The rates of glucose and lipid oxidation were calculated from the gas exchange measurements by using nonprotein values to the following equations (31): lipid (milligrams per minute) = (1.6946 × VO₂ − 1.7012 × VCO₂)/BW (kilograms); and glucose (milligrams per minute) = (4.585 × VCO₂ − 3.2255 × VO₂)/BW (kilograms).

The power calculations for achieving significance for the main outcomes (muscle and liver insulin sensitivity) of the insulin clamp studies were calculated a priori using a coefficient of variation estimated to be between 5 and 9% based on all of our previous insulin clamp studies. To further limit variations in outcomes, the subjects were very carefully matched for age (young); gender; physical activity; lifestyle; nonsmoking status; freedom from medical conditions; body mass, BMI, and fat mass; and part of the menstrual cycle. Furthermore, all subjects underwent careful study preparations including dietary and activity instructions and were admitted to the research unit in the evening before the insulin clamp studies.

Statistical analyses were performed using Stat View (Abacus Concepts, Berkeley, CA). To detect statistical differences between control and LBW subjects, unpaired Student t tests were performed for independent samples with a P < 0.05 considered significant. Nonnormally distributed data (i.e. area under the curve data) were log transformed. All data are expressed as mean ± sem.

Results

The control and LBW groups (n = 45 each) were matched for age, BMI, ethnicity, and physical activity; despite this matching, the LBW group was on average 3.0 ± 0.1 cm shorter in stature than the control group (Table 1). All subjects had a sedentary lifestyle and did not participate in any exercise regimens as documented by the activity questionnaire and 3 consecutive days of ad libitum activity monitoring using pedometers.

Table 1.

Characteristics of the entire cohorts of control and LBW subjects

	Control (10 males/35 females)	LBW (10 males/35 females)	P value
Age (yr)	24 ± 1	24 ± 1	0.97
Ethnicity (Asian/Black/Caucasian/Hispanic)	5 A/5 B/31 C/3 H	3 A/6 B/31 C/5 H
Height (m)	1.67 ± 0.01	1.64 ± 0.01	0.07
Body weight (kg)	61.7 ± 1.3	61.6 ± 1.6	0.94
BMI (kg/m2)	22.1 ± 0.3	22.9 ± 0.5	0.19
Birth weight (kg)	3.4 ± 0.1	2.0 ± 0.1	<0.0001
Fasting plasma glucose (mg/dl)	91.3 ± 1.2	88.7 ± 1.2	0.112
Fasting plasma insulin (μU/ml)	8.2 ± 0.4	11.4 ± 7.8	0.011
HOMA	2.16 ± 0.18	2.49 ± 0.18	0.19
ISI (10−4 dl/min per mU/ml)	5.57 ± 0.23	4.32 ± 0.23	0.0017
IMCL (%)a	0.91 ± 0.05	0.89 ± 0.06	0.77
EMCL (%)a	1.05 ± 0.07	1.08 ± 0.12	0.83
Liver TG (%)b	0.58 ± 0.09	0.70 ± 0.11	0.44
Activity index	2.23 ± 0.11	2.14 ± 0.12	0.60
Activity (miles/d)	3.36 ± 0.20	3.41 ± 0.31	0.89

A, Asian; B, Black; C, Caucasian; H, Hispanic; HOMA, homeostasis model assessment; EMCL, extramyocellular lipid; TG, triglyceride.

^an = 44 controls, n = 42 LBW individuals.

^bn = 44 controls, n = 41 LBW individuals.

All subjects had normal fasting glucose concentrations, normal glucose tolerance, and similar area under the curve (AUC) for glucose during the OGTT (Fig. 1A). Fasting plasma insulin concentrations were 40% higher in the LBW subjects compared with the control subjects (P = 0.011) (Table 1) and remained significantly higher during the OGTT than plasma insulin concentrations in the control group. The AUC for insulin was 40% higher in the LBW than in the control group (Fig. 1B) (P = 0.0001). The AUC for C-peptide was 22% higher in the LBW compared with the control group, although this was not different (P = 0.222) (Fig. 1C). The ISI, which reflects whole-body insulin sensitivity, was 20% lower in the LBW group than in the control group (P = 0.0017) (Table 1).

Fig. 1.

Time courses for plasma concentrations during the 3-h OGTT. A, Glucose (milligrams per deciliter), AUC for glucose (P = NS). B, Insulin (microunits per milliliter), AUC for insulin (P = 0.0001). C, C-peptide concentrations (nanomoles per liter), AUC for C-peptide (P = 0.09).

Fasting plasma concentrations of GH were approximately 200% higher (P = 0.01) and fasting plasma cortisol concentrations were 27% higher (P = 0.01) in the LBW groups compared with the control group (Table 2). Fasting plasma concentrations of total IGF-I were 19% lower in the LBW compared with the control subjects (P < 0.05). In contrast, there were no differences in plasma concentrations of IGF-I BP-3, adiponectin, IL-6, or TNF-α between the control and LBW groups (Table 2).

Table 2.

Fasting plasma concentrations of GH, cortisol, total IGF-I, IGF-I BP-3, adiponectin, IL-6, and TNF-α

	Control (n = 45)	LBW (n = 45)	P value
GH (ng/ml)	3.8 ± 0.8	7.4 ± 1.1	0.01
Cortisol (μg/dl)	17.1 ± 1.0	21.7 ± 1.5	0.01
Total IGF-I (ng/ml)	383 ± 30	309 ± 22	<0.05
BP-3 (ng/ml)a	4845 ± 129	4744 ± 177	0.65
Adiponectin (μg/ml)	10.6 ± 0.6	12.1 ± 1.0	0.18
IL-6a	1.34 ± 0.11	2.05 ± 0.38	0.37
TNF-αa	1.47 ± 0.13	2.28 ± 0.92	0.60

^an = 39 per group.

The IMCL, extramyocellular lipid, and liver triglyceride content, as determined by ¹H MRS, were similar in the LBW (n = 41) and control (n = 43) groups (Table 1).

Muscle insulin sensitivity was examined in subgroups of age-, gender-, BMI-, percent fat mass-, and activity-matched control (n = 3 males and n = 7 females; age 28 ± 2 yr; BMI 22.1 ± 1.0 kg/m²; body fat 24.9 ± 1.3%) and LBW subjects (n = 3 males and n = 7 females; age 29 ± 2 yr; BMI 22.5 ± 0.5 kg/m²; body fat 25.0 ± 1.9%) during a euglycemic-hyperinsulinemic clamp of 20 mU/(m²·min). Whole-body insulin resistance in the LBW subjects was confirmed by a 61% reduction in rates of insulin stimulated glucose infusion during the clamp compared with the control subjects (P = 0.0003). Under these conditions hepatic glucose production was completely suppressed in both groups, whereas insulin stimulated muscle glucose uptake was 62% lower in the LBW group than in the control (P = 0.0003) (Fig. 2). This muscle insulin resistance in the LBW group could be attributed mostly to a 92% reduction in nonoxidative glucose disposal (P = 0.00005) (Table 3). These differences in glucose uptake between the groups were similar when expressed per lean body mass.

Fig. 2.

Rates of insulin-stimulated glucose metabolism in control (n = 10) and LBW (n = 10) subjects during a euglycemic-hyperinsulinemic [20 mU/(m²·min)] clamp.

Table 3.

Basal and insulin-stimulated respiratory quotients, energy expenditure, and whole-body glucose and lipid oxidation as determined by indirect calorimetry

	Control (n = 10)	LBW (n = 10)	P value
Basal respiratory quotient	0.79 ± 0.02	0.83 ± 0.01	0.12
Clamp respiratory quotient	0.87 ± 0.03	0.87 ± 0.01	0.97
Basal energy expenditure (kcal/kg BW per 24 h)	22.7 ± 0.9	24.7 ± 0.9	0.142
Clamp energy expenditure (kcal/ kg BW per 24 h)	23.2 ± 0.9	23.7 ± 0.8	0.692
Rates of basal glucose oxidation (mg/kg BW per minute)	1.26 ± 0.29	01.78 ± 0.11	0.131
Rates of clamp glucose oxidation (mg/kg BW per minute)	2.58 ± 0.44	2.50 ± 0.16	0.866
Rates of clamp nonoxidative glucose metabolism (mg/kg BW per minute)	5.13 ± 0.69	0.42 ± 0.17	0.00005
Basal lipid oxidation (mg/kg BW per minute)	1.20 ± 0.11	1.11 ± 0.06	0.519
Clamp lipid oxidation (mg/kg BW per minute)	0.74 ± 0.15	0.78 ± 0.09	0.821

There were no differences in basal or insulin clamp energy expenditure or respiratory quotient values between the groups (Table 3).

There were no significant differences in fasting or insulin-stimulated rates of whole-body glucose or fat oxidation between the two groups (Table 3).

Hepatic insulin sensitivity was specifically assessed in a second hyperinsulinemic-euglycemic clamp study of 10 mU/(m²·min) in subgroups of control (n = 2 males and n = 3 females; age 27 ± 5 yr; BMI 22.8 ± 0.4 kg/m²; body fat 21.6 ± 5.6%) and LBW subjects (n = 2 males and n = 3 females; age 26 ± 2 yr; BMI 21.3 ± 0.8 kg/m²; body fat 19.4 ± 4.0%) (n = 5) matched for age, gender, BMI, percent body fat, and physical activity. Rates of fasting glucose production were similar in the two groups but during this low-dose insulin infusion additional hepatic insulin resistance was uncovered in the LBW group as reflected by an impairment of insulin to suppress glucose production (48 ± 7%) compared with the control group (86 ± 9%; P = 0.01) (Fig. 3).

Fig. 3.

Percent suppression of hepatic glucose production in control (n = 5) and LBW (n = 5) subjects during a hyperinsulinemic-euglycemic [10 mU/(m²·min)] clamp.

Rates of [²H₅] glycerol turnover during the baseline and both of the euglycemic-hyperinsulinemic clamps were similar in the control and LBW groups and the percent insulin suppression of glycerol turnover was similar in the groups. However, this was not a primary end point of this study, and the number of subjects may be inadequate to detect differences between the groups. Rates of [²H₅] glycerol turnover during the baseline and the 20 mU/(m²·min) euglycemic-hyperinsulinemic clamp are shown in Fig. 4.

Discussion

The role of insulin resistance in the pathogenesis of T2D in LBW individuals remains controversial, with some studies finding that LBW individuals are insulin resistant (5, 34–36), whereas others found that they were not (7, 8, 10). Furthermore, most of the studies that found insulin resistance in LBW subjects did not identify the organ(s) or define the mechanism responsible for the insulin resistance. These discrepancies may be due to lack of careful control for key variables that are known to affect insulin sensitivity such as activity, body weight, age, medications, and smoking. Furthermore, insulin doses used may have been too high to reveal insulin resistance in both muscle and liver (23). To circumvent these potential limitations, we examined liver, muscle, and adipose tissue insulin responsiveness as well as insulin secretion in healthy young, lean, sedentary, nonsmoking LBW individuals and age-, weight-, and activity-matched, normal-birth-weight individuals. Using this approach, we found that LBW individuals manifested liver and muscle insulin resistance.

Previous studies by our group (11–14, 18, 20, 22) and others (11, 23) have demonstrated a strong relationship between liver and muscle insulin resistance and ectopic lipid accumulation in these tissues. In this study of LBW individuals, we found no relationship between liver and muscle insulin resistance and IMCL and liver lipid accumulation in these organs. This study is the first, to our knowledge, to disassociate liver and muscle insulin resistance from ectopic lipid accumulation in liver and muscle in healthy, normal-weight LBW adults, suggesting that some other mechanism is responsible for the insulin resistance in LBW individuals. In this regard we found that liver and muscle insulin resistance in LBW individuals was associated with significant increases in fasting plasma cortisol and GH concentrations and reductions in plasma IGF-I concentrations. Furthermore, the lower plasma IGF-I concentrations were associated with a 3-cm average reduction in stature in the LBW individuals. The alterations in fasting plasma cortisol, GH, and IGF-I concentrations, although representing only a single time point, are in accordance with a previous study, which found increased plasma cortisol levels in LBW children accompanied by low plasma IGF-I concentrations and short stature (37). Consistent with these observations, clinical studies of preeclampsia and intrauterine growth retardation, conditions known to cause LBW in humans, have also found elevated glucocorticoid concentrations in umbilical cord blood samples (6, 9). Further studies using more detailed measures to assess the HPA and GH/IGF-I axes in LBW individuals will be of interest.

Studies of insulin secretion and β-cell function in LBW subjects have provided conflicting results, with some studies suggesting impaired β-cell insulin secretion in LBW subjects as a potential reason or their increased prevalence of T2D (5, 8), whereas other studies have found insulin secretion to be normal (38–41) or increased in LBW subjects (42). In the current study, we did not find any evidence of impaired insulin secretion, as assessed by frequent blood sampling during the OGTT. In contrast, AUC for both insulin and C-peptide concentrations were higher in the LBW than in the control group, which is consistent with our findings of muscle and liver insulin resistance in LBW individuals.

In conclusion, we found that healthy young lean LBW individuals manifest liver and muscle insulin resistance, which was disassociated from ectopic lipid accumulation in these organs. In contrast, we found the LBW subjects to have increased plasma concentrations of GH and cortisol and reductions in plasma IGF-I concentrations. Taken together, these observations support the hypothesis that alterations in the regulation of the hypothalamic-pituitary axis may be a contributing factor to the liver and muscle insulin resistance in LBW individuals and a contributing factor to their increased prevalence of T2D.