The effects of pregnancy and type 1 diabetes [insulin-dependent diabetes mellitus (IDDM)] on protein metabolism are still uncertain. Therefore, six normal and five IDDM women were studied during and after pregnancy, using [13C]leucine and [2H5]phenylalanine with a hyperinsulinemic-euglycemic clamp and amino acid infusion. Fasting total plasma amino acids were lower in pregnancy in normal but not IDDM women (2,631 ± 427 vs. 2,057 ± 471 and 2,523 ± 430 vs. 2,500 ± 440 μmol/l, respectively). Whole body protein breakdown (leucine) increased in pregnancy [change in normal (ΔN) and IDDM women (ΔD) 0.59 ± 0.40 and 0.48 ± 0.26 g · kg−1 · day−1, bothP < 0.001], whereas reductions in protein breakdown due to insulin/amino acids (ΔN −0.57 ± 0.19, ΔD −0.58 ± 0.20 g · kg−1 · day−1, both P < 0.001) were unaffected by pregnancy. Protein breakdown in IDDM women was not higher than normal, and neither pregnancy nor type 1 diabetes altered the insulin sensitivity of amino acid turnover. Nonoxidized leucine disposal (protein synthesis) increased in pregnancy (ΔN 0.67 ± 0.45, ΔD 0.64 ± 0.34 g · kg−1 · day−1, bothP < 0.001). Pregnancy reduced the response of phenylalanine hydroxylation to insulin/amino acids in both groups (ΔN −1.14 ± 0.74, ΔD −1.12 ± 0.77 g · kg−1 · day−1, bothP < 0.05). These alterations may enable amino acid conservation for protein synthesis and accretion in late pregnancy. Well-controlled type 1 diabetes caused no abnormalities in the regulation of basal or stimulated protein metabolism.
- isotope labeling
- glucose clamp technique
- insulin resistance
- insulin-dependent diabetes mellitus
protein metabolism in human pregnancy has been evaluated using primed constant infusions of stable isotope-labeled amino acids such as [13C]leucine and [2H5]phenylalanine (17, 24, 51). These cross-sectional studies in the fasted state have shown contradictory results with large variations within study groups, finding protein breakdown decreased (17) or increased (51) and protein synthesis decreased (17) or unchanged (51). Studies in nonpregnant subjects have shown that insulin in combination with amino acid infusion promotes protein deposition by reducing protein breakdown and stimulating protein synthesis and oxidation (4, 11, 29, 47). We proposed that the augmentation of protein accretion by pregnancy would occur during insulin/amino acid infusion, and, in the absence of previous studies, we have assessed protein metabolism with serial insulin/amino acid infusions during and after pregnancy in a group of normal subjects.
Studies of protein metabolism in nonpregnant type 1 diabetes [insulin-dependent diabetes mellitus (IDDM)] using both fasted and insulin- or amino acid-stimulated conditions have shown the loss of insulin's inhibitory effect on protein breakdown and oxidation (6, 23, 29, 38, 48, 49, 54). This may be exacerbated during pregnancy and particularly in the postprandial state. A study of protein metabolism in the fasted state (24) found that the pregnant IDDM subjects had higher rates of protein breakdown and oxidation but similar protein synthesis rates compared with normal pregnant subjects. The resulting increase in plasma amino acids could mean greater substrate availability for placental transfer and fetal metabolism and growth. To assess this possible mechanism for the development of fetal macrosomia, we have examined whether IDDM affects protein metabolism in pregnancy with reduced insulin sensitivity of protein breakdown.
l-[1-13C]leucine (99%13C),l-[ring-2H5]phenylalanine (98% 2H),l-[ring-2H4]tyrosine (98% 2H), and sodium [13C]bicarbonate (99%13C) were purchased from Cambridge Isotopes Laboratories (Woburn, MA). They were dissolved in normal saline and tested to be sterile and pyrogen free.
Six healthy normal pregnant women (without medical problems or medication) and six women with IDDM were recruited from the antenatal clinics and joint obstetrics/medical clinic at the Royal Victoria Infirmary. One IDDM woman did not return for postpartum study and was excluded from analysis. IDDM women had all been receiving insulin for at least a year before becoming pregnant. Duration of diabetes was 1, 6, 14, 18, and 22 yr. Basal serum C-peptide levels in late pregnancy were 0.14 nmol/l in one and <0.05 nmol/l in the other four diabetic subjects. All IDDM women were deemed to have good glycemic control at their late pregnancy study (Hb A1c 6.6 ± 0.3%, mean ± SD). This pregnancy was the first child for four normal and two IDDM women and the second child for two normal and three IDDM women. All except one subject produced healthy live offspring without obstetric complications. One of the IDDM subjects had an intrapartum fetal death due to placental abruption in labor. Postmortem examination showed an otherwise full-term healthy infant. Informed consent was obtained from all the women, and ethical approval was given by the Joint Ethics Committee of our Health Authority and University. All of the studies were conducted in late pregnancy (34–38 wk gestation) and were repeated at least 12 wk postpartum when breast feeding had ceased and before oral contraception had been resumed.
Each subject attended the research unit after an overnight fast, having kept a protein diet of 65 g/day for 3 days before the study. On arrival, each subject voided her bladder and had her height and weight measured. Bioelectric impedance was measured with a meter (Holtain) using standardized procedures, and percent lean body mass was calculated using an equation evaluated in pregnant women (55). This was followed by cannulation of both hands, one for infusion of the stable isotopes, amino acid mixture, glucose, and insulin and the other for sampling arterialized superficial venous blood by the hot-box technique.
IDDM subjects omitted their morning insulin dose, and their fasting blood glucose was checked before the start of the tracer infusions. If their blood glucose was between 4 and 5 mmol/l, the main infusions were begun. However, if their blood glucose was >5 mmol/l, they were first given intravenous insulin to bring their blood glucose levels to the target range before commencement of the tracer amino acids. The insulin infusion was stopped once the target blood glucose level was reached. During the initial 3 h of the study (i.e., the basal state, 0–180 min), an intravenous infusion of 20% dextrose was kept on standby to be infused as required if two consecutive blood glucose readings were <4 mmol/l. The dextrose infusion was given at a variable rate or was discontinued to maintain blood glucose within the target range. Four out of the five IDDM subjects had fasting blood glucose >5 mmol/l and required insulin at both study visits to reduce their blood glucose to the target range before isotope infusion. The mean time interval required was 1 h, with 3–8 mU insulin required during pregnancy and 2–16 mU when not pregnant (equivalent to ∼0.1–0.2 mU · kg−1 · h−1). No further insulin dose was required on commencement of the isotope infusions.
The subjects were given priming doses of [13C]leucine (0.5 mg/kg), [2H5]phenylalanine (0.5 mg/kg), [2H4]tyrosine (0.08 mg/kg), and sodium [13C]bicarbonate (0.08 mg/kg) followed immediately by continuous infusions of [13C]leucine (0.5 mg · kg−1 · h−1) and [2H5]phenylalanine (0.5 mg · kg−1 · h−1) for 6 h. After 3 h, a hyperinsulinemic-euglycemic clamp was added together with an infusion of a mixture of amino acids (Vamin 14). Blood and expired breath samples were collected at −30, −15, and 0 min (averaged as initial) before the start of the infusions and thereafter at 120, 150, 160, 170, and 180 min (averaged as basal) and 300, 330, 340, 350, and 360 min (averaged as insulin/amino acid infusion). Steady-state conditions were checked by the lack of significant slope using a linear regression of enrichment values against time. Indirect calorimetry measurements for 20 min with a Deltatrac metabolic monitor were carried out before the start and again in the last 30 min of the basal and insulin/amino acid infusion periods.
During the hyperinsulinemic-euglycemic clamp, subjects were clamped with a target blood glucose concentration of 4.5 mmol/l with a variable-rate 20% dextrose infusion at the same time as the insulin infusion. Blood glucose was measured using a glucose analyzer (Yellow Springs Instruments) every 10 min throughout the clamp period (this 10-min monitoring was maintained throughout the whole study period for the IDDM subjects). All subjects received 40 mU · kg−1 · h−1 insulin infusion prepared by drawing up the required amount of human Actrapid monocomponent insulin (100 IU/ml; Novo-Nordisk) mixed with normal saline to a volume of 24 ml in a 30-ml syringe; the insulin was delivered by a syringe pump at a rate of 6 ml/h. The glucose used for infusion was derived from potato starch, which is equivalent in13C content to normal dietary intake in the United Kingdom (43); the potato starch was obtained from Avebe (Humberside, UK). This was made up in 500-ml bottles of 20% dextrose suitable for infusion in humans, and 20 mmol of potassium chloride were preadded to the dextrose infusion at the time of manufacture to prevent hypokalemia. Electrolyte-free Vamin 14 (Pharmacia) was infused with a priming dose of 52 mg/kg amino acid over 10 min followed by a continuous infusion at a rate of 0.52 ml · kg−1 · h−1, which provided 23.8 μmol · kg−1 · h−1 of leucine and 18.9 μmol · kg−1 · h−1 of phenylalanine. Plasma amino acid levels were at steady state after one half hour [data from pilot studies (unpublished) and similar published reports (4)]. The average anabolic stimulus included ∼5 g of amino acids and 18 g of glucose in the final hour.
All samples were centrifuged at 4°C immediately after collection, and the plasma was stored at −70°C until analysis. The plasma samples were derivatized using a combination of methods (9, 20). The α-ketoisocaproic acid (KIC) was derivatized to its quinoxalinol-tert-butyldimethylsilyl derivative. Phenylalanine and tyrosine were derivatized to theirtert-butyldimethylsilyl derivatives. Electron ionization gas chromatography-mass spectrometry (GC-MS) analysis was performed on the derivatives separately using a Finnegan 1020 GC-MS to determine the13C or deuterium enrichments using selected-ion monitoring. Phenylalanine was measured at mass-to-charge ratio (m/z) 336 and 341 and tyrosine atm/z 466, 468, and 470. Appropriate corrections were made for all the samples using the calibration equations derived from standards for each amino acid. When only [2H4]tyrosine was infused, enrichments of [2H2]tyrosine remained at basal levels, confirming no degradation of label during analysis (52). Breath 13CO2 enrichment was measured by isotope-ratio mass spectrometry (ANCA; Europa Scientific).
Models for Protein Metabolism
Whole body protein metabolism is commonly assessed from essential amino acid kinetics, and leucine and phenylalanine were used simultaneously as tracers since they represent two independent methods of assessing whole body protein metabolism. Leucine is catabolized mainly in muscle (2), and phenylalanine is catabolized in liver (53).
Leucine flux was determined by measuring the 13C enrichment of arterialized plasma KIC, which provides a well-accepted estimate of the intracellular enrichment of leucine (3, 31) and the precursor pools for protein synthesis and leucine catabolism. Leucine oxidation was calculated from breath 13CO2 and plasma [13C]KIC enrichment. Nonoxidized leucine disposal (hereafter called protein synthesis) was calculated by subtracting leucine oxidation from leucine flux. We have used values of 0.74 and 0.86 (nonpregnant, basal and insulin/amino acid infusion) and 0.86 and 0.99 (pregnant, basal and insulin/amino acid infusion) for bicarbonate recovery based on previous studies on bicarbonate recovery in pregnant women and clamp conditions in our unit (13, 40). The leucine content of whole body protein was assumed to be 629 μmol/g protein (35).
The phenylalanine model has been described (52). The calculation of amino acid flux used the standard equation. Our calculation of phenylalanine hydroxylation in pregnancy used our value of 0.52 for the molar ratio of the fluxes of tyrosine and phenylalanine arising from protein breakdown; the value was previously derived from the infusion of [2H2]tyrosine in pregnant women (57). This enabled a simplified infusion protocol that otherwise was made awkward by the limited solubility of [2H2]tyrosine. However, because of the unavailability of a plasma metabolite to indicate intracellular precursor enrichment, correction factors are recommended (39) to permit a reasonable equivalence of data from [2H5]phenylalanine and [13C]phenylalanine methods (30). Phenylalanine flux was divided by 0.691 (analogous to the KIC-to-leucine ratio), and phenylalanine hydroxylation was multiplied by 1.45 (fasted) and 1.78 (insulin/amino acid infusion), derived from estimates of hepatic precursor pool enrichments during apolipoprotein B-100 synthesis (39). Given the uncertainty of these calculations and their extrapolation to pregnancy, it has been suggested that estimation of protein synthesis rates may be unwise (33), and so only breakdown and phenylalanine hydroxylation (before and after correction) are reported. Whole body metabolism in grams per kilogram per day was calculated by assuming the phenylalanine content of whole body protein to be 280.3 μmol/g protein (35).
The two-tailed Student's t-test for paired data was used within normal and IDDM groups to compare data from basal vs. insulin/amino acid periods or from the pregnant vs. nonpregnant occasions. These significances are shown in Tables 1-7. Two-way ANOVA with repeated measures (using GENSTAT 5 and MANOVA; SPSS) was used to analyze the effects and interactions of two factors (e.g., insulin/amino acid infusion and pregnancy or pregnancy and diabetes) on a variable (e.g., leucine breakdown). Significance was taken as 5% or less (8).
The clinical details of the subjects are shown in Table1. Lean body mass (%) did not vary significantly between late pregnancy and 12 wk postpartum, as seen previously (56), and so subsequent results were expressed per kilogram body weight. The time course of KIC and phenylalanine enrichments (Fig. 1) shows the subjects to be at steady state in the last hour of the basal and insulin/amino acid infusion periods. The mean percent enrichments at steady state for [13C]KIC, [2H5]phenylalanine, [2H4]tyrosine, and13CO2 are shown in Table2. The mean enrichments of the amino acid tracers were lower in pregnancy than in the nonpregnant state, significantly so for tyrosine. Under basal conditions, [2H4]tyrosine at steady state showed a difference in enrichment due to pregnancy of 0.76 ± 0.64 (P < 0.001, n = 15), whereas [2H5]phenylalanine had a difference in enrichment of 0.2 ± 0.72 (not significant). The coefficient of variation (CV) at steady state was between 3.5 and 6.3% for [13C]KIC, [2H5]phenylalanine, and 13CO2. The CV at steady state for [2H4]tyrosine was higher between 7.8 and 10.5%. There were no significant differences in enrichments between subject groups.
The progress of the hyperinsulinemic-euglycemic clamps is shown in Fig.2. In the last hour of the clamp, normal subjects achieved a blood glucose of 4.5 ± 0.1 mmol/l with a glucose infusion rate (GIR) of 3.0 ± 0.1 mg · kg−1 · min−1 when pregnant and a blood glucose of 4.5 ± 0.1 mmol/l with a GIR of 4.8 ± 0.4 mg · kg−1 · min−1 when not pregnant. IDDM subjects achieved a blood glucose of 4.7 ± 0.2 mmol/l with a GIR of 3.8 ± 0.2 mg · kg−1 · min−1 when pregnant and a glucose of 4.7 ± 0.1 mmol/l with a GIR of 3.9 ± 0.3 mg · kg−1 · min−1 when not pregnant.
For both subject groups, Table 3 shows whole body protein breakdown, synthesis, oxidation, and net balance derived from leucine. Table 4 shows protein breakdown and hydroxylation data derived from phenylalanine. Tables 3 and 4 also show the effect of insulin/amino acid infusion and pregnancy on protein metabolism with significance levels from pairedt-tests. Results below give the combined significance level from two-way ANOVA.
Insulin/amino acid infusion caused a reduction in protein breakdown (P = 0.026) and a rise in protein oxidation (P < 0.001) but had little effect on protein synthesis. Protein balance was improved by insulin/amino acid infusion, especially when pregnant (P < 0.001).
In pregnancy, protein breakdown was higher compared with the nonpregnant state (P = 0.024). Protein synthesis was also higher during pregnancy (P = 0.009). Mean protein oxidation was unchanged by pregnancy.
In response to insulin/amino acid infusion, there was a reduction in protein breakdown (P = 0.003). Protein oxidation increased with insulin/amino acid infusion (P = 0.001), and this was greater in the nonpregnant state (significant interaction between insulin/amino acid infusion and pregnancy, P = 0.012). When pregnant, the stimulation of oxidation by insulin/amino acid infusion was less than in the normal subjects (P = 0.03). Insulin/amino acid infusion had little effect on protein synthesis. The improvement in protein balance due to insulin/amino acid infusion was doubled during pregnancy compared with the nonpregnant state (P = 0.001).
In pregnancy, protein breakdown was higher than when not pregnant (P = 0.009). Protein synthesis was higher during pregnancy (P = 0.002). Pregnancy reduced (P = 0.013) the response of protein oxidation to the insulin/amino acid infusion (significant interaction, P= 0.012). (Pregnancy did not affect oxidation in normal subjects.) Pregnancy had a positive improvement on protein balance during insulin/amino acid infusion.
None of the significance of the changes described here was reduced by the correction factors described in methods. Insulin/amino acid infusion reduced protein breakdown (P = 0.026) and increased phenylalanine hydroxylation (P < 0.001). Pregnancy caused a reduction in the response of phenylalanine hydroxylation to insulin/amino acid infusion (−1.14 ± 0.74 g · kg−1 · day−1,P < 0.05).
In pregnancy, mean protein breakdown was unchanged compared with the nonpregnant state. Phenylalanine hydroxylation was markedly reduced during pregnancy (P < 0.001), especially during the insulin/amino acid infusion (interaction P = 0.018).
During insulin/amino acid infusion, there was no change in mean protein breakdown. (In normal subjects, we had seen that insulin/amino acid infusion caused breakdown to fall.) Insulin/amino acid infusion caused a rise in phenylalanine hydroxylation (P < 0.001). Pregnancy caused a reduction in the response of phenylalanine hydroxylation to insulin/amino acid infusion (−1.12 ± 0.77 g · kg−1 · day−1,P < 0.05).
During pregnancy, protein breakdown was higher (P = 0.005), whereas normal subjects showed no response. When pregnant, phenylalanine hydroxylation was reduced (P < 0.001), especially during the insulin/amino acid infusion (interactionP = 0.035).
Tables 5 and6 show the mean changes in leucine, phenylalanine, and total amino acids throughout the 6-h study period. Fasting total plasma amino acids were lower in pregnancy in normal but not IDDM women. For both groups, there was little change in plasma leucine concentration during the insulin/amino acid infusion, whereas phenylalanine concentration doubled. Mean initial and basal-stage insulin levels were at least doubled during pregnancy compared with nonpregnancy for the IDDM subjects, whereas levels for the insulin clamp period were similar. In normal subjects, insulin levels showed no difference between pregnancy and nonpregnancy.
Table 7 shows the individual amino acid levels at the end of the basal and insulin/amino acid infusion stages of the protocol. In the basal state during pregnancy, our IDDM subjects had significantly higher (P < 0.05) levels of alanine, glutamic acid, histidine, methionine, and threonine. Glycine, isoleucine, and leucine were significantly lower (P < 0.05) during pregnancy for our IDDM subjects. Valine concentrations were also reduced but did not reach statistical significance.
Plasma leucine and phenylalanine kinetics were studied serially in six normal and five IDDM women during and after pregnancy using hyperinsulinemic-euglycemic clamps with concurrent amino acid infusion. Previous studies of protein metabolism in pregnancy used the fasted state and were cross-sectional. We proposed that the alterations of protein metabolism by pregnancy and IDDM would be more clearly revealed by serial studies using anabolic stimuli with subjects as their own controls. The uncertainties in calculation of phenylalanine and tyrosine metabolic parameters have already been alluded to inmethods. We therefore have concentrated thediscussion on analysis of the leucine data and the effect of insulin/amino acid infusion, pregnancy, and diabetes.
Insulin/Amino Acid Infusion in Normal Subjects
During anabolic stimulation, there was a reduction in protein breakdown and a rise in protein oxidation whether subjects were pregnant or nonpregnant. This concurred with other studies in nonpregnant subjects (4, 11, 23, 29, 47). There have been no other studies of protein metabolism during insulin/amino acid infusion in pregnant subjects. Our data on protein synthesis during insulin/amino acid infusion showed no response in either pregnant or nonpregnant states. Studies in nonpregnant subjects using leucine have found a rise in protein synthesis in response to amino acids alone (5, 22, 37), whereas others found no change (14,32). Insulin alone reduced protein synthesis (11,29) or had no significant effect (16, 48, 50, 54). Studies of insulin/amino acid infusion have all shown a rise in protein synthesis (4, 11, 23, 29, 47). Tessari et al. (45) concluded that hyperaminoacidemia was more important than insulin in stimulating protein synthesis in vivo. Insulin increases tissue uptake of amino acids, especially branched-chain amino acids like leucine, (19) to muscle and causes a reduction in circulating amino acid concentrations (7), including leucine. Our amino acid infusion provided greater amounts of leucine than phenylalanine, and increased uptake in the presence of insulin may account for the unchanged plasma leucine levels during the insulin/amino acid infusion (Table 5) while phenylalanine concentrations doubled. When we infused Vamin 14 without the insulin clamp during a separate validation study (unpublished observation), leucine levels did rise (67–101 μmol/l). The infusion rate of leucine that we used with the clamp increased substrate availability of leucine from the tissue free amino acid pool but led to increased oxidation and reduced breakdown rather than synthesis. Infusion of four times the rate of leucine we used achieved a significant increase in plasma leucine levels and a rise in protein synthesis from leucine (6, 23). However, the intramuscular leucine concentration still remained unchanged because of the large rise in oxidation (6). This diversion of intracellular substrate availability has been put forward as an explanation for the lack of increase in whole body protein synthesis during insulin/amino acid infusion with the leucine model (34). Insulin/amino acid infusion did improve protein balance, and this was achieved mainly through decreased protein breakdown. Net balance during insulin/amino acid infusion was approximately zero during pregnancy. This hyperinsulinemic-euglycemic model routinely does not add extra energy apart from the requirements to maintain euglycemia (∼18 g of glucose in the final hour), and recent studies have shown net protein balance to be barely positive, even when plasma leucine levels doubled (46).
Insulin/Amino Acid Infusion in IDDM Subjects
During anabolic stimulation, we found a significant reduction in protein breakdown with the leucine model whether pregnant or not pregnant. Studies on protein metabolism in uncontrolled IDDM using leucine have shown that the inhibitory effect of insulin on protein breakdown and oxidation was missing (44). In both the basal and insulin/amino acid-infused states there was increased breakdown, raised plasma leucine levels, and increased rates of leucine oxidation, whereas protein synthesis was unchanged or increased. Correction of glycemic control with insulin normalized protein metabolism (27, 36) but with greater peripheral free insulin levels. Almost all changes in whole body protein synthesis in IDDM due to insulin treatment occurred in the splanchic region (1). There has been no comparable study in pregnant IDDM subjects. We found no overall response of leucine-derived synthesis to insulin/amino acid infusion. A lack of stimulation of skeletal protein synthesis with insulin/amino acid infusion using the leucine model in IDDM subjects has been shown (6). Bennet et al. (6) suggested a possible postreceptor defect of protein metabolism, making the insulin ineffective in IDDM subjects. However, another study using leucine and phenylalanine tracers in IDDM subjects with insulin/amino acid infusion concluded that there was no impairment of leucine utilization for protein synthesis in IDDM subjects (23). We found that, although protein oxidation was increased with insulin/amino acid infusion, protein balance improved mainly by a reduction in breakdown as in normal subjects.
Effect of Pregnancy in Normal Subjects
In pregnancy, there was increased protein breakdown measured by the leucine model. Previous studies in late pregnancy, using leucine in the fasted state, found breakdown decreased (17) or increased (51). End-product analysis, using glycine, found no difference between early and late pregnancy, but this model could not differentiate between fed and fasted states (18). Increased protein breakdown during the fasted state and a normal response to insulin may account for higher turnover seen during the insulin clamp rather than the suggestion of decreased amino acid sensitivity to insulin in late gestation (12). Greater amino acid availability in late pregnancy may be part of the compensatory mechanism to replenish nutrient losses to the fetus from maternal plasma and indirectly supports the “accelerated starvation” theory (21). Previous studies in late pregnancy, using leucine in the fasted state, found synthesis decreased (17) or unchanged (51). We found that protein synthesis was increased in pregnancy. The leucine model showed little change in protein oxidation during pregnancy, as seen in previous studies (13, 51). Pregnancy had little effect on net protein balance (with or without insulin/amino acid infusion), largely because the increase in protein synthesis was offset by an increase in breakdown.
Analysis of the effect of pregnancy is complicated by the accumulation of increased body mass (∼10 kg in mother, fetus, and placenta) that, while having similar lean body composition, may have different metabolic activity. It is relevant to consider the possible impact of fetal metabolism on the assessment of maternal protein metabolism. Studies of fetal metabolism have shown protein breakdown to be resistant to suppression by insulin, and enhanced protein accretion was secondary to decreased leucine oxidation (28). Studies of neonates suggested that the enhanced stimulation of skeletal muscle protein by feeding was mediated by insulin, whereas stimulation of liver protein synthesis was a function of amino acid supply (15).
Pregnancy Effect in IDDM Subjects
In pregnancy, there was increased protein breakdown despite higher mean fasting plasma levels of insulin compared with pregnant normal subjects (17 ± 10 vs. 7 ± 5 mU/l, respectively), which would be expected to diminish protein breakdown. Protein breakdown in the fasted state has been found to be normal (59) or higher in pregnant IDDM subjects (24) than in normal subjects, suggesting that there could be diabetes-associated resistance to the effects of insulin not only on carbohydrate but also on protein metabolism. However, our findings of similar reductions in leucine breakdown during insulin/amino acid infusion whether pregnant or not pregnant, normal or diabetic, suggested no short-term resistance to exogenous insulin in respect of protein metabolism. Protein synthesis was increased in pregnancy in IDDM subjects, and the increase in protein synthesis due to pregnancy was greater during insulin/amino acid infusion compared with the postabsorptive state. The rise in protein oxidation during insulin/amino acid infusion was reduced by one-half during pregnancy and suggested that altered sensitivity to anabolic stimulation enabled increased conservation of amino acids for protein synthesis and deposition in pregnancy, achieving positive protein balance.
Leucine vs. Phenylalanine Models
In normal subjects, both models showed that insulin/amino acid infusion decreased protein breakdown and increased protein oxidation. In our IDDM subjects, the two amino acid models showed that protein breakdown was increased in pregnancy and that the rise in protein oxidation during insulin/amino acid infusion was reduced in pregnancy. Previous comparisons of the leucine and phenylalanine methods in male IDDM subjects concluded that both methods generated similar protein metabolism results (20, 23). One would expect the responses of protein breakdown to be similar with both models if they accurately reflect whole body protein metabolism. Differences in protein oxidation rates may be related to differences at sites of amino acid catabolism, since leucine oxidation to KIC occurs in muscle, whereas phenylalanine hydroxylation to tyrosine occurs predominantly in liver. Differences in plasma availability may also be responsible (46).
Comparisons Between the Normal and IDDM Groups
ANOVA showed that there were no significant differences in protein breakdown and synthesis in the basal or insulin/amino acid infusion state, during or after pregnancy. None of the protein metabolism parameters showed any significant interaction between the response to pregnancy and diabetes.
It was hypothesized that leucine-derived protein breakdown in response to insulin/amino acid infusion and pregnancy would be higher in IDDM subjects. We found no significant interaction of diabetes in the response to insulin/amino acid infusion. Some studies have suggested that intensive glycemic control in IDDM pregnancies led to normalization of other metabolic fuels (41). In IDDM subjects, correction of glycemic control with insulin resulted in normal protein metabolism parameters (44). However, IDDM subjects regarded by current tools as being well controlled still have major abnormalities, including peripheral hyperinsulinemia, abnormal plasma insulin profiles, and marked postprandial hyperglycemia (26). Our protein metabolism results for pregnant IDDM subjects in the basal state were in contrast to others (24) who found (using leucine) higher protein breakdown and oxidation rates in the fasted state in pregnant IDDM subjects despite good glycemic control.
Total amino acid concentrations showed a pregnancy-associated fall in normal subjects as noted by us previously (58) but no fall in our IDDM subjects, becoming elevated in pregnancy compared with normal subjects. This was particularly so after insulin/amino acid infusion. The higher levels of gluconeogenic amino acids in our IDDM subjects may be due to reduced utilization for gluconeogenesis because of higher insulin concentrations in the IDDM subjects. The branched-chain amino acids may also be reduced because of higher insulin concentrations. Increased skeletal muscle release of amino acids has been suggested as the cause of possible elevated total amino acid concentrations in pregnant IDDM subjects because of pregnancy-associated insulin resistance (25). Thus the higher maternal fasting amino acid levels would allow greater transplacental transfer, potentially driving the tendency to macrosomia, whereas postprandial regulation allows normal maintenance of maternal protein balance in IDDM subjects. Protein breakdown was increased during pregnancy in both normal and IDDM subjects despite higher basal insulin concentrations in the latter group, and there was a normal response of protein breakdown to exogenous insulin during pregnancy by IDDM subjects. Despite this, in our clinic population with similar blood glucose control to the IDDM study group, almost 50% of subjects have macrosomic neonates (10). The lack of a relationship between the degree of blood glucose control and macrosomia in well-controlled IDDM women has been observed by us (10) and others (42).
In summary, our original findings were that the effect in pregnancy of the insulin/amino acid infusion was consistent with that observed in nonpregnant subjects by other investigators. The effect of pregnancy was to increase rates of protein breakdown and protein synthesis, whereas leucine oxidation was unaltered. Phenylalanine hydroxylation was much lower during pregnancy. These alterations would enable amino acid conservation for protein synthesis and accretion in late pregnancy. The presence of IDDM caused no significant difference in protein breakdown, synthesis, or oxidation between the normal and IDDM subjects. This is in contrast to previous studies suggesting reduced insulin sensitivity of protein breakdown and oxidation as a potential cause of macrosomia.
We are grateful to David Halliday and Brendan Cooper for technical advice and statistical assistance and to Dr. John Matthews, Dept. of Medical Statistics, Newcastle University, for advice.
This work was supported by the British Diabetic Association.
Address for reprint requests and other correspondence: P. G. Whittaker, 7905 Winston Rd., Philadelphia, PA 19118 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2000 the American Physiological Society