You are here

Article Text

Article menu
Pathophysiology/complications
Decreased branched-chain amino acids and elevated fatty acids during antecedent hypoglycemia in type 1 diabetes
  1. Rui She1,2,
  2. Naba Hassan Al-sari3,
  3. Ismo Matias Mattila3,
  4. Anne-Sophie Sejling4,
  5. Jens Pedersen5,
  6. Cristina Legido-Quigley3,6,
  7. Ulrik Pedersen-Bjergaard1,2
  1. 1Department of Endocrinology and Nephrology, Nordsjællands Hospital, Hillerød, Capital Region of Denmark, Denmark
  2. 2Department of Clinical Medicine, University of Copenhagen Faculty of Health and Medical Sciences, Kobenhavn, Capital Region of Denmark, Denmark
  3. 3Clinical Research, Steno Diabetes Center Copenhagen, Herlev, Capital Region of Denmark, Denmark
  4. 4Boston Global Development, Novo Nordisk, Søborg, Capital Region of Denmark, Denmark
  5. 5Department of Internal Medicine, Herlev Hospital, Herlev, Capital Region of Denmark, Denmark
  6. 6Institute of Pharmaceutical Science, King's College London, London, UK
  1. Correspondence to Ulrik Pedersen-Bjergaard; ulrik.pedersen-bjergaard{at}regionh.dk

Abstract

Introduction Hypoglycemia is a major limiting factor in achieving recommended glycemic targets for people with type 1 diabetes. Exposure to recurrent hypoglycemia results in blunted hormonal counter-regulatory and symptomatic responses to hypoglycemia. Limited data on metabolic adaptation to recurrent hypoglycemia are available. This study examined the acute metabolic responses to hypoglycemia and the effect of antecedent hypoglycemia on these responses in type 1 diabetes.

Research design and methods Twenty-one outpatients with type 1 diabetes with normal or impaired awareness of hypoglycemia participated in a study assessing the response to hypoglycemia on 2 consecutive days by a hyperinsulinemic glucose clamp. Participants underwent a period of normoglycemia and a period of hypoglycemia during the hyperinsulinemic glucose clamp. Plasma samples were taken during normoglycemia and at the beginning and the end of the hypoglycemic period. Metabolomic analysis of the plasma samples was conducted using comprehensive two-dimensional gas chromatography with time-of-flight mass spectrometry.

Results In total, 68 metabolites were studied. On day 1, concentrations of the branched-chain amino acids, leucine (p=3.8×10−3) and isoleucine (p=2.2×10−3), decreased during hypoglycemia. On day 2, during hypoglycemia, five amino acids (including leucine and isoleucine) significantly decreased, and two fatty acids (tetradecanoic and oleic acids) significantly increased (p<0.05). Although more metabolites responded to hypoglycemia on day 2, the responses of the single metabolites were not statistically significant between the 2 days.

Conclusions In individuals with type 1 diabetes, one episode of hypoglycemia decreases leucine and isoleucine concentrations. Antecedent hypoglycemia results in the decrement of five amino acids and increases the concentrations of two fatty acids, suggesting an alteration between the two hypoglycemic episodes, which could indicate a possible adaptation. However, more studies are needed to gain a comprehensive understanding of the consequences of these alterations.

Trial registration number NCT01337362.

  • Type 1 diabetes mellitus
  • Diabetes complication
  • Hypoglycemia
  • Energy metabolism

Data availability statement

Data are available upon reasonable request.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

https://doi.org/10.1136/bmjdrc-2023-003327

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    WHAT IS ALREADY KNOWN ON THIS TOPIC

    • Hypoglycemia is a major limiting factor for obtaining optimal glycemic control in individuals with insulin-treated diabetes.

    • Recurrent hypoglycemia attenuates the counter-regulatory hormonal response, which is responsible for the clinical symptoms and metabolic response to hypoglycemia.

    WHAT THIS STUDY ADDS

    • This study suggests that antecedent hypoglycemia enhances the metabolic response to a subsequent hypoglycemic episode, indicating a compensatory metabolic response to the attenuation of the counter-regulatory hormonal response.

    HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

    • The metabolic alterations to recurrent hypoglycemia might be further explored in mechanistic studies and may potentially be a target for intervention.

    Introduction

    Hypoglycemia is the most common adverse event of insulin replacement therapy in type 1 diabetes (T1D).1 2 To recover from hypoglycemia and avoid developing more severe hypoglycemia, people with diabetes depend on hormonal counter-regulatory responses that increase glycogenolysis and gluconeogenesis and the generation of warning symptoms that prompt carbohydrate ingestion. However, these responses tend to diminish with the increasing duration of diabetes.

    Studies investigating the impact of hypoglycemia in individuals with T1D and healthy people have shown that antecedent hypoglycemic episodes lead to blunted counter-regulatory and symptomatic responses, indicating that short-term adaptational processes occur. In T1D, the glucagon response to hypoglycemia becomes grossly diminished in most patients after a few years.3 Furthermore, many individuals with T1D develop counter-regulatory failure with attenuated centrally mediated autonomic responses to hypoglycemia resulting in decreased epinephrine secretion, subsequent impairment of the glucose counter-regulation and blunting of autonomic warning symptoms, which over time leads to impaired awareness of hypoglycemia.1 During hypoglycemia, the decrease in plasma glucose levels and elevation in counter-regulatory hormones change energy metabolism.4 In healthy human subjects, the plasma concentrations of all amino acids are reduced5 while the plasma concentration of free fatty acids is elevated, supporting lipolysis and stimulating gluconeogenesis.6 7 In T1D, the metabolic response to hypoglycemia is much less understood. One study found that in T1D, isoleucine and leucine were suppressed during hypoglycemia but significantly less suppressed compared with healthy controls.8 Another found that individuals with T1D seemed to preserve their lipid response to hypoglycemia compared with healthy controls.9 These relatively small studies have given us insights into the metabolic responses; however, the area remains largely unexplored.

    Metabolomic profiling has provided the possibility to obtain detailed metabolic insights in various conditions. In diabetes, previous studies have primarily focused on risk profiling and preconditional metabolic changes.10–12 One study of insulin-induced hypoglycemia in type 2 diabetes used metabolomics and found a change in more than 90 metabolites during hypoglycemia that was almost absent after 24 hours.13 The metabolic response to hypoglycemia in people with T1D using metabolomics as a platform has, to our knowledge, not previously been reported.

    This study aims to assess metabolic responses to a single episode of hyperinsulinemic hypoglycemia in individuals with T1D and the short-term metabolic adaptation to a subsequent hypoglycemic episode 24 hours later. We also examine whether the state of hypoglycemia awareness is associated with these responses.

    Methods

    The study is a post hoc analysis of data from an experimental mechanistic clinical study investigating electroencephalogram changes and cognitive function during antecedent hypoglycemia in participants with T1D and either hypoglycemia awareness or unawareness.14 The protocol is registered at http://clinicaltrials.gov (NCT01337362). All participants were informed about the study procedures and provided written consent to participate.

    Recruitment

    Participants, both male and female over 18 years old, with T1D for over 5 years and with normal hypoglycemia awareness or unawareness were included. Participants were recruited from the outpatient clinics at Nordsjællands Hospital and Steno Diabetes Center Copenhagen, Denmark. Individuals with pregnancy, brain disorders and cardiovascular disease were excluded from the study. Other exclusion criteria included antiepileptic, β-blocker drug use, and alcohol and drug abuse. We included all the participants we acquired blood sampling from at all three time points during both experimental days (figure 1).

    Figure 1
    Figure 1

    Timeline of the clamp set-up during day 1 and day 2. A hypoglycemic glycemic clamp was performed under normoglycemic and hypoglycemic levels. The hyperinsulinemic-normoglycemic period started when patients during the clamp reached plasma glucose between 5 and 6 mmol/L. Plasma glucose was measured every 10 min, and blood samples (BS) for counter-regulatory hormones and metabolomic analysis were drawn three times at the end of the normoglycemic period, the beginning of the hypoglycemic period and the end of the hypoglycemic period.

    Hypoglycemia awareness status

    Hypoglycemia awareness status was classified according to three validated methods: the Pedersen-Bjergaard method,15 the Gold score16 and the Clarke method.17 Participants were classified as having normal hypoglycemia awareness if they scored within the normal awareness category on all three methods. Participants were classified as hypoglycemia unaware if they scored within the unaware category using the Pedersen-Bjergaard method, impaired using the Gold score and reduced using the Clarke method. Participants who did not meet the criteria for normal hypoglycemia awareness or hypoglycemia unawareness were excluded from the study.

    Experimental design

    Five days before the planned experiment, every participant was equipped with a continuous glucose monitor (CGM) (Guardian Real-Time with Enlite sensor; Medtronic, Minneapolis). Participants used a blood glucose meter (Contour Link; Bayer HealthCare, Leverkusen, Germany) to monitor and calibrate the CGM. Each participant arrived fasted at the research unit on the morning of their scheduled appointment. The experiment was postponed for 2 weeks if CGM showed glucose measurements below 3.5 mmol/L 24 hours prior to the first day. If no hypoglycemia was detected, the participant underwent a hyperinsulinemic glucose clamp in normoglycemia and hypoglycemia, acquiring plasma blood samples throughout the experiment. The following day, the same experimental procedure was performed again.

    Hyperinsulinemic glucose clamp procedure

    Insulin and glucose were infused intravenously during the entire experiment. Human insulin (Actrapid, Novo Nordisk, Ballerup, Denmark) was infused with a continuous rate of 1 mU insulin/(kg⋅min). Initially, a 20% glucose solution was infused at the same rate for all participants (0.015 mmol/(kg⋅min)) but thereafter adjusted to reach the desired blood glucose target. The blood glucose level target was 5.0–6.0 mmol/L during normoglycemia and recovery from hypoglycemia, while the target was 2.0–2.5 mmol/L during hypoglycemia. Participants were held at each target period for 1 hour (figure 1).

    Blood sampling and biochemical analyses

    Plasma glucose was continuously analyzed using YSI 2300 (YSI/Xylem, Yellow Springs, Ohio), and blood was sampled from the cubital vein. Furthermore, additional blood samples were acquired three times during the experiment: at the end of the normoglycemic period, the beginning of the hypoglycemic period, and the end of the hypoglycemic period. Blood samples were used to measure the counter-regulatory hormones: glucagon, epinephrine, norepinephrine, cortisol, growth hormone and substrates. Data on the hormonal counter-regulatory responses have previously been presented by Sejling et al.14 Additionally, blood samples for metabolomic analysis were acquired.

    Metabolomic analysis

    Sample preparation and analysis were performed as described before.18 Briefly, plasma samples (30 μL) were spiked with 400 μL methanol and 10 μL of an internal standard mixture (d4-succinic acid, d5-glutamic acid, d8-valine, and d33-heptadecanoic acid; Sigma Aldrich). Samples were vortex mixed and incubated on ice for 30 min and centrifuged (10 000 rpm, 3 min, 4°C). Finally, 180 µL of the filtered extracts was transferred to glass vials and evaporated to dryness before derivatization. The samples were derivatized using a previously described procedure.19 Briefly, derivatization converts the reactive biological groups into trimethylsilyl derivatives, increasing the biomolecules’ volatility. The polar metabolites were analyzed using a Pegasus 4D (LECO; Saint Joseph, USA) system, which combines two-dimensional chromatographic separation with time-of-flight mass spectrometric detection.

    The data files obtained by the ChromaTOF software were exported to text files. In-house developed software Guineu20 was used for aligning the data for further analyses. The original two-dimensional gas chromatography–time-of-flight mass spectrometry (GC×GC–TOFMS) data include retention times, retention indices (RI), spectral information for possible identification, spectral similarity value (S=0–999), and peak response data. The detected peaks were aligned using the Guineu software.20 The alignment of the data was performed based on RI, second dimension retention times and spectra. Identifications were assigned using the National Institute of Standards and Technology and in-house libraries. Data were then postprocessed and analyzed in R (https://www.r-project.org/).21

    Statistical analysis

    Data were analyzed in R. Metabolite-wise mixed-effects models were used to consider the six repeated measures from each participant in the trial. Statistical tests were corrected for multiple testing and comparison with false discovery rate (FDR) using the Benjamini-Hochberg method.22

    First, differences between the 2 days were modeled with metabolite-wise mixed-effects models, with time, day and time-day interaction as fixed effects and the participant ID as random effect. We further conducted the metabolite concentration changes found from the mentioned analysis within days. These changes were visualized in a heatmap. In the heatmap, each colored rectangle corresponds to the median concentration of the metabolite levels that were p value adjusted <0.05. Metabolites with the most substantial differences each day were visualized with curves, where the observations were grouped by time point. Finally, differences between aware and unaware groups were explored. A detailed description of the data analysis plan is available in online supplemental material 1.

    Supplemental material

    Results

    Baseline characteristics of the participants

    A total of 21 individuals with T1D, both female and male, were included in the study. The participants had a body mass index mean value of 24 kg/m2 and a relatively long duration of disease (28 years). HbA1c mean value was 8.0% for the participants, and they used an average daily insulin dose of 41 IU. Eleven of the 21 participants were hypoglycemia unaware (table 1).

    View this table:
    Table 1

    Baseline characteristics

    Metabolomics

    Sixty-eight metabolites from the plasma samples were identified and semiquantified using untargeted GC×GC–TOFMS metabolomic analysis.

    Metabolic responses on day 1

    In total, two out of 68 metabolites had a change in their levels during day 1; the concentrations of two branched-chain amino acids (BCAA), isoleucine and leucine, were significantly decreased as a response to hypoglycemia compared with baseline (p<0.05, figure 2A); isoleucine (β±SE: −0.72±0.16, p=2.2×10−3), leucine (β±SE: −0.78±0.18, p=3.8×10−3). Another 10 amino acids were identified in the analysis shown in table 2. None of the other amino acids had a significant change; however, all showed a decrease in tendency. Metabolite names, 95% CI, FDR p values and slopes are given in table 2. The complete table is available in online supplemental material 2.

    Supplemental material

    Figure 2
    Figure 2

    Metabolic responses of amino acids and free fatty acids on day 1 (A) and day 2 (B, C). SEM error bars. Samples were acquired at the end of the normoglycemic period (Normo), the beginning of the hypoglycemic period (Hypo1), and the end of the hypoglycemic period (Hypo2). Day 1: only leucine and isoleucine had significantly different levels between the end of the normoglycemic period and at the end of the hypoglycemic period (p<0.05). Day 2: five amino acids and two fatty acids showed significantly different levels between the end of the normoglycemic period and at the end of the hypoglycemic period (p<0.05).

    View this table:
    Table 2

    Results of the mixed-effects model

    Metabolic responses on day 2

    During day 2, the concentration of seven metabolites (including those from day 1) was significantly changed; five amino acids (leucine, valine, isoleucine, methionine, and phenylalanine) were significantly decreased (p<0.05), and two fatty acids (tetradecanoic acid (myristic acid 14:0) and oleic acid) were increased (figure 2B, C). The eight other amino acids did not show significant changes as seen on day 1; however, they still had a decreasing trend. Metabolite names, 95% CI, FDR p values and slopes are given in table 3. The complete table is available in online supplemental material 3.

    Supplemental material

    View this table:
    Table 3

    Results of the mixed-effects model

    Adaptation and impact of hypoglycemia awareness status

    Seven metabolites responded to hypoglycemia on day 2 compared with two metabolites on day 1. Metabolites responding to hypoglycemia on both days followed the same trend. However, no statistically significant differences were found when comparing the metabolic responses to hypoglycemia between the days. Additionally, no significant difference between the metabolic responses in individuals with normal or impaired awareness could be detected before and after correcting for multiple testing (p>0.94). Complete data sets of the comparative analysis between normal and impaired awareness are available in online supplemental materials 4 and 5.

    Supplemental material

    Discussion

    This untargeted metabolic profiling of individuals with T1D under a hyperinsulinemic-normoglycemic and hypoglycemic clamp condition showed dynamic changes in several metabolites that increased following antecedent hypoglycemia. On day 1, we found that an event of hypoglycemia decreased the BCAAs: isoleucine and leucine. A subsequent hypoglycemic event 24 hours later resulted in a decrease in additional three amino acids (phenylalanine, valine, and methionine) and an increase in two fatty acids (myristic and oleic acids). Despite this suggestion of an adaptation toward a more dynamic metabolic response to hypoglycemia on day 2, there were no statistically significant differences in the responses of the single metabolites between the 2 days. Interestingly, in this study, alanine, an important energy fuel for muscle and other tissues,23 is unaltered during hypoglycemia on both days. However, our study cannot exclude other possibilities; one suggestion could be that BCAAs are converted into alanine, maintaining plasma alanine concentration while also being a factor in the decrement of the plasma BCAA levels.24

    The changes in amino acids correspond to the findings in healthy human studies, which have shown a reduction in BCAA concentration during acute hyperinsulinemic hypoglycemia.5 25 Thus, the decrease in plasma amino acids could reflect increased utilization of amino acids as substrates for ketone bodies and gluconeogenesis.26 Battezzati et al have shown that the decrease in plasma concentrations of leucine and phenylalanine was significantly smaller when comparing individuals with T1D and healthy controls,8 suggesting an impairment of this counter-regulatory response in T1D. Our study is not suited to support their findings; however, exploring this would be very interesting. Furthermore, the metabolomics platform employed in our study could only detect 12 out of the 20 standard proteinogenic amino acids, as shown in tables 2 and 3. Therefore, future studies using metabolomics should use platforms capable of identifying all amino acids. Doing so would provide a more comprehensive understanding of the alterations in amino acid levels associated with hypoglycemia.

    On day 2, alongside the downregulation of five amino acids, we also observed increased levels of two fatty acids, tetradecanoic acid and oleic acid, which could suggest an increase in lipolysis. However, in our study, glycerol, another product of lipolysis, did not change in plasma levels during hypoglycemia.27 Therefore, our study cannot determine whether the elevation in Free Fatty Acids (FFAs) results from increased lipolysis or suppression of the utilization of FFAs, which in both cases would lead to an increase in plasma FFA. However, one study examining lipolysis and hypoglycemia in T1D found an elevation in lipolysis and an increase in FFAs,9 indicating that our findings could also result from increased lipolysis. Additionally, our data explored the fingerprint of the changes in FFA during hypoglycemia and showed a faster response in specific FFAs, oleic acid and tetradecanoic acid.

    FFAs are important substrates during fasted conditions. FFAs are substrates for hepatic β oxidation providing energy in the form of ATP and forming ketone bodies which can be transported to other extrahepatic tissues via the circulation.28 Although β oxidation does not produce any substrates for gluconeogenesis, the ATP formed by β oxidation is used in facilitating gluconeogenesis.28 Our study found that 3-hydroxybutyric acid (3-OHB), a ketone body, was unaltered during hypoglycemia on both days (tables 2 and 3), indicating that ketogenesis does not seem to increase during hypoglycemia in our study. Another study examining healthy participants during hypoglycemia used a bolus insulin injection to induce either hypoglycemia or normoglycemia randomly and found that while FFA increased immediately after inducing hypoglycemia, 3-OHB levels from the hypoglycemia-induced participants did not increase until plasma insulin levels were down to baseline.29 In contrast, we used a continuous insulin infusion which probably severely suppressed ketogenesis.

    The metabolic responses to hypoglycemia are primarily stimulated by the counter-regulatory hormonal response.30 As mentioned, the hormonal responses in this study have previously been presented by Sejling et al.14 Briefly, all counter-regulatory hormones (glucagon, epinephrine, cortisol and growth hormone) increased on both days. There were no differences between the 2 days for the hormonal responses except for growth hormone, which was lower during the second day. In T1D, the glucagon response to glucagon is severely diminished,31 while growth hormone and cortisol do not seem to have an effect during the first hour of hypoglycemia.32 33 These factors make epinephrine the pivotal counter-regulatory hormonal response during the acute phase. Previous studies investigating epinephrine infusion in healthy subjects and patients with T1D have shown a similar fall in amino acid plasma levels.34 35 Although we did not find a statistically significant difference in amino acids between the 2 study days, we did observe more amino acids downregulated on day 2 than on day 1. This alteration in metabolic response could be a result of a change in the cellular response to epinephrine since the counter-regulatory response from epinephrine did not differ between the first and subsequent episodes of hypoglycemia.14

    Previous studies have shown diverging results regarding the physiological responses to epinephrine. Guy et al found that the physiological response to epinephrine is altered in T1D, finding lower glucose levels and cardiovascular responses. However, they did see a greater increase in lipolytic response compared with healthy subjects.36 Hypoglycemia also seems to impact the physiological response of epinephrine, reducing the cardiovascular, hepatic, and adipose tissue responses to epinephrine.37 Another study, however, did not see the same changes.38 All mentioned studies did not investigate the changes in different metabolites as we did; however, due to the limitation of our study, which did not specifically focus on metabolic responses to epinephrine, further studies looking at the amino acid and lipid fluxes are needed to elaborate our findings.

    Although the glucagon response to hypoglycemia is deficient in T1D,31 we did observe an increase in our study.14 Previous studies investigating the effects of glucagon on protein metabolism in healthy subjects have shown that glucagon may lower leucine concentration but not plasma concentration of phenylalanine.39 Another study, however, did not find a significant alteration in the plasma concentration of BCAAs in response to glucagon.40 Thus, it is still debatable whether the changes we see in the plasma concentrations of the BCAAs are influenced by glucagon. In our study, this is further supported by the minimal glucagon responses.14

    Another important hormone is insulin, which alters protein metabolism, lowers the plasma concentration of amino acids41 42 and stimulates lipogenesis, thus decreasing the plasma concentration of fatty acids.43 In our study, however, baseline normoglycemic plasma samples and plasma taken in hypoglycemia were the same during a constant hyperinsulinemic conditions, only altering the glucose infusion. Thus, the observed metabolic changes should not be a result of insulin. However, as discussed, the continuous insulin infusion in our study seems to suppress some metabolic pathways and to fully understand the metabolic consequences of the alterations, future studies aimed at investigating how these changes could facilitate energy metabolism should include experiments using bolus insulin to induce hypoglycemia, thus limiting the influence of the continuous hyperinsulinemic condition found in a clamp study.

    There were no differences in the metabolic reactions of people with normal hypoglycemia awareness and unawareness on either of the days. This could be explained by the indifferent hormonal response between the two groups, reported previously by Sejling et al.14 Similar findings were also observed in a previous study investigating metabolic responses to hypoglycemia in T1D.9

    In conclusion, this study shows a significant metabolic response to hypoglycemia in people with long-standing T1D and indicates that recurrent hypoglycemia alters the metabolic response, particularly within protein and lipid metabolism. Our results could indicate that while the counter-regulatory hormonal response attenuates to recurrent hypoglycemia, the metabolism may try to adapt to these conditions. However, future studies are needed to confirm and elaborate our findings and explore the consequences. The strengths of this study include a considerable sample size and a rigorous clamp protocol. However, to further understand metabolic responses to hypoglycemia in T1D, patients with newly onset diabetes duration and healthy controls could have been included.

    Supplemental material

    Data availability statement

    Data are available upon reasonable request.

    Ethics statements

    Patient consent for publication

    Ethics approval

    This study involves human participants and was approved by the Regional Committee on Health Research Ethics (ID H-1-2011-024). Participants gave informed consent to participate in the study before taking part.

    Acknowledgments

    The authors thank the study participants and research nurses P Banck-Petersen, T Larsen and C Hansen from the Department of Endocrinology and Nephrology at Nordsjællands Hospital Hillerød for technical assistance in conducting the clinical trial. The authors also thank and acknowledge the laboratory assistance by the biomedical scientists at Steno Diabetes Center Copenhagen.

    References

    1. ↵
      1. Cryer PE
      . Banting lecture. Hypoglycemia: the limiting factor in the management of IDDM. Diabetes 1994;43:1378–89. doi:10.2337/diab.43.11.1378
      OpenUrlAbstract/FREE Full Text
    2. ↵
      1. Geller AI,
      2. Shehab N,
      3. Lovegrove MC, et al
      . National estimates of insulin-related hypoglycemia and errors leading to emergency department visits and hospitalizations. JAMA Intern Med 2014;174:678–86. doi:10.1001/jamainternmed.2014.136
      OpenUrl
    3. ↵
      1. Arbelaez AM,
      2. Xing D,
      3. Cryer PE, et al
      . Blunted glucagon but not epinephrine responses to hypoglycemia occurs in youth with less than 1 yr duration of type 1 diabetes mellitus. Pediatr Diabetes 2014;15:127–34. doi:10.1111/pedi.12070
      OpenUrlCrossRefPubMed
    4. ↵
      1. Fanelli C,
      2. Pampanelli S,
      3. Epifano L, et al
      . Relative roles of insulin and hypoglycaemia on induction of neuroendocrine responses to, symptoms of, and deterioration of cognitive function in hypoglycaemia in male and female humans. Diabetologia 1994;37:797–807. doi:10.1007/BF00404337
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      1. De Feo P,
      2. Perriello G,
      3. Santeusanio F, et al
      . Differential effects of insulin-induced Hypoglycaemia on the plasma branched-chain and non-branched-chain amino acid concentrations in humans. Diabete Metab 1992;18:277–82.
      OpenUrlPubMed
    6. ↵
      1. Caprio S,
      2. Gelfand RA,
      3. Tamborlane WV, et al
      . Oxidative fuel metabolism during mild hypoglycemia: critical role of free fatty acids. Am J Physiol 1989;256:E413–9. doi:10.1152/ajpendo.1989.256.3.E413
      OpenUrl
    7. ↵
      1. Fanelli C,
      2. Calderone S,
      3. Epifano L, et al
      . Demonstration of a critical role for free fatty acids in mediating counterregulatory stimulation of gluconeogenesis and suppression of glucose utilization in humans. J Clin Invest 1993;92:1617–22. doi:10.1172/JCI116746
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      1. Battezzati A,
      2. Benedini S,
      3. Sereni LP, et al
      . Protein and glutamine kinetics during counter-regulatory failure in type 1 diabetes. Nutr Metab Cardiovasc Dis 2009;19:352–7. doi:10.1016/j.numecd.2008.03.009
      OpenUrlPubMed
    9. ↵
      1. Bengtsen MB,
      2. Støy J,
      3. Rittig NF, et al
      . A human randomized controlled trial comparing metabolic responses to single and repeated hypoglycemia in type 1 diabetes. J Clin Endocrinol Metab 2020;105:dgaa645. doi:10.1210/clinem/dgaa645
    10. ↵
      1. Frohnert BI,
      2. Webb-Robertson B-J,
      3. Bramer LM, et al
      . Predictive modeling of type 1 diabetes stages using disparate data sources. Diabetes 2020;69:238–48. doi:10.2337/db18-1263
      OpenUrlAbstract/FREE Full Text
    11. ↵
      1. Tofte N,
      2. Suvitaival T,
      3. Trost K, et al
      . Metabolomic assessment reveals alteration in polyols and branched chain amino acids associated with present and future renal impairment in a discovery cohort of 637 persons with type 1 diabetes. Front Endocrinol (Lausanne) 2019;10:818. doi:10.3389/fendo.2019.00818
      OpenUrl
    12. ↵
      1. Wang TJ,
      2. Larson MG,
      3. Vasan RS, et al
      . Metabolite profiles and the risk of developing diabetes. Nat Med 2011;17:448–53. doi:10.1038/nm.2307
      OpenUrlCrossRefPubMedWeb of Science
    13. ↵
      1. Halama A,
      2. Kahal H,
      3. Bhagwat AM, et al
      . Metabolic and proteomic signatures of hypoglycaemia in type 2 diabetes. Diabetes Obes Metab 2019;21:909–19. doi:10.1111/dom.13602
      OpenUrl
    14. ↵
      1. Sejling A-S,
      2. Kjaer TW,
      3. Pedersen-Bjergaard U, et al
      . Hypoglycemia-associated EEG changes following antecedent hypoglycemia in type 1 diabetes mellitus. Diabetes Technol Ther 2017;19:85–90. doi:10.1089/dia.2016.0331
      OpenUrl
    15. ↵
      1. Pedersen-Bjergaard U,
      2. Pramming S,
      3. Thorsteinsson B
      . Recall of severe hypoglycaemia and self-estimated state of awareness in type 1 diabetes. Diabetes Metab Res Rev 2003;19:232–40. doi:10.1002/dmrr.377
      OpenUrlCrossRefPubMedWeb of Science
    16. ↵
      1. Gold AE,
      2. MacLeod KM,
      3. Frier BM
      . Frequency of severe hypoglycemia in patients with type I diabetes with impaired awareness of hypoglycemia. Diabetes Care 1994;17:697–703. doi:10.2337/diacare.17.7.697
      OpenUrlAbstract/FREE Full Text
    17. ↵
      1. Clarke WL,
      2. Cox DJ,
      3. Gonder-Frederick LA, et al
      . Reduced awareness of hypoglycemia in adults with IDDM. A prospective study of hypoglycemic frequency and associated symptoms. Diabetes Care 1995;18:517–22. doi:10.2337/diacare.18.4.517
      OpenUrlAbstract/FREE Full Text
    18. ↵
      1. Orešič M,
      2. Anderson G,
      3. Mattila I, et al
      . Targeted serum metabolite profiling identifies metabolic signatures in patients with Alzheimer’s disease, normal pressure hydrocephalus and brain tumor. Front Neurosci 2017;11:747. doi:10.3389/fnins.2017.00747
    19. ↵
      1. Gika HG,
      2. Theodoridis GA,
      3. Plumb RS, et al
      . Current practice of liquid chromatography-mass spectrometry in metabolomics and metabonomics. J Pharm Biomed Anal 2014;87:12–25. doi:10.1016/j.jpba.2013.06.032
      OpenUrlCrossRefPubMed
    20. ↵
      1. Castillo S,
      2. Mattila I,
      3. Miettinen J, et al
      . Data analysis tool for comprehensive two-dimensional gas chromatography/time-of-flight mass spectrometry. Anal Chem 2011;83:3058–67. doi:10.1021/ac103308x
      OpenUrlCrossRefPubMed
    21. ↵
      The R project for statistical computing; 2019.
    22. ↵
      1. Storey JD,
      2. Tibshirani R
      . Statistical significance for genomewide studies. Proc Natl Acad Sci U S A 2003;100:9440–5. doi:10.1073/pnas.1530509100
      OpenUrlAbstract/FREE Full Text
    23. ↵
      1. Felig P
      . The glucose-alanine cycle. Metabolism 1973;22:179–207. doi:10.1016/0026-0495(73)90269-2
      OpenUrlCrossRefPubMedWeb of Science
    24. ↵
      1. Holecek M
      . Relation between glutamine, branched-chain amino acids, and protein metabolism. Nutrition 2002;18:130–3. doi:10.1016/s0899-9007(01)00767-5
      OpenUrlCrossRefPubMedWeb of Science
    25. ↵
      1. Schauder P,
      2. Schröder K,
      3. Matthaei D, et al
      . Influence of insulin on blood levels of branched chain keto and amino acids in man. Metabolism 1983;32:323–7. doi:10.1016/0026-0495(83)90038-0
      OpenUrlPubMed
    26. ↵
      1. Berg JM,
      2. Stryer L,
      3. Tymoczko J, et al
      . Biochemistry. 9th edn. New York: W.H. Freeman/Macmillan Learning, 2019: xlii,1096, 45, 60, 45.
    27. ↵
      1. Schweiger M,
      2. Eichmann TO,
      3. Taschler U, et al
      . Measurement of lipolysis. Methods Enzymol 2014;538:171–93. doi:10.1016/B978-0-12-800280-3.00010-4
      OpenUrlCrossRefPubMed
    28. ↵
      1. Rui L
      . Energy metabolism in the liver. Compr Physiol 2014;4:177–97. doi:10.1002/cphy.c130024
      OpenUrlPubMed
    29. ↵
      1. Voss TS,
      2. Vendelbo MH,
      3. Kampmann U, et al
      . Effects of insulin-induced hypoglycaemia on lipolysis rate, lipid oxidation and adipose tissue signalling in human volunteers: a randomised clinical study. Diabetologia 2017;60:143–52. doi:10.1007/s00125-016-4126-x
      OpenUrl
    30. ↵
      1. Cryer PE
      . Mechanisms of hypoglycemia-associated autonomic failure in diabetes. N Engl J Med 2013;369:362–72. doi:10.1056/NEJMra1215228
      OpenUrlCrossRefPubMedWeb of Science
    31. ↵
      1. Bisgaard Bengtsen M,
      2. Møller N
      . Mini-review: glucagon responses in type 1 diabetes - a matter of complexity. Physiol Rep 2021;9:e15009. doi:10.14814/phy2.15009
    32. ↵
      1. Møller N,
      2. Jørgensen JOL
      . Effects of growth hormone on glucose, lipid, and protein metabolism in human subjects. Endocr Rev 2009;30:152–77. doi:10.1210/er.2008-0027
      OpenUrlCrossRefPubMedWeb of Science
    33. ↵
      1. Stanley S,
      2. Moheet A,
      3. Seaquist ER
      . Central mechanisms of glucose sensing and counterregulation in defense of hypoglycemia. Endocr Rev 2019;40:768–88. doi:10.1210/er.2018-00226
      OpenUrlCrossRef
    34. ↵
      1. Shamoon H,
      2. Jacob R,
      3. Sherwin RS
      . Epinephrine-induced hypoaminoacidemia in normal and diabetic human subjects: effect of beta blockade. Diabetes 1980;29:875–81. doi:10.2337/diab.29.11.875
      OpenUrlAbstract/FREE Full Text
    35. ↵
      1. Matthews DE,
      2. Pesola G,
      3. Campbell RG
      . Effect of epinephrine on amino acid and energy metabolism in humans. Am J Physiol 1990;258:E948–56. doi:10.1152/ajpendo.1990.258.6.E948
      OpenUrl
    36. ↵
      1. Guy DA,
      2. Sandoval D,
      3. Richardson MA, et al
      . Differing physiological effects of epinephrine in type 1 diabetes and nondiabetic humans. Am J Physiol Endocrinol Metab 2005;288:E178–86. doi:10.1152/ajpendo.00310.2004
      OpenUrl
    37. ↵
      1. Aftab Guy D,
      2. Sandoval D,
      3. Richardson MA, et al
      . Effects of glycemic control on target organ responses to epinephrine in type 1 diabetes. Am J Physiol Endocrinol Metab 2005;289:E258–65. doi:10.1152/ajpendo.00311.2004
      OpenUrlCrossRefPubMed
    38. ↵
      1. Hoffman RP
      . Antecedent hypoglycemia does not alter increased epinephrine-induced lipolysis in type 1 diabetes mellitus. Metabolism 2006;55:371–80. doi:10.1016/j.metabol.2005.09.012
      OpenUrlPubMed
    39. ↵
      1. Battezzati A,
      2. Simonson DC,
      3. Luzi L, et al
      . Glucagon increases glutamine uptake without affecting glutamine release in humans. Metabolism 1998;47:713–23. doi:10.1016/s0026-0495(98)90036-1
      OpenUrlCrossRefPubMedWeb of Science
    40. ↵
      1. Couet C,
      2. Fukagawa NK,
      3. Matthews DE, et al
      . Plasma amino acid kinetics during acute states of glucagon deficiency and excess in healthy adults. Am J Physiol 1990;258:E78–85. doi:10.1152/ajpendo.1990.258.1.E78
      OpenUrl
    41. ↵
      1. Nair KS,
      2. Ford GC,
      3. Ekberg K, et al
      . Protein dynamics in whole body and in splanchnic and leg tissues in type I diabetic patients. J Clin Invest 1995;95:2926–37. doi:10.1172/JCI118000
      OpenUrlCrossRefPubMedWeb of Science
    42. ↵
      1. Nair KS,
      2. Ford GC,
      3. Halliday D
      . Effect of intravenous insulin treatment on in vivo whole body leucine kinetics and oxygen consumption in insulin-deprived type I diabetic patients. Metabolism 1987;36:491–5. doi:10.1016/0026-0495(87)90049-7
      OpenUrlCrossRefPubMedWeb of Science
    43. ↵
      1. Dimitriadis G,
      2. Mitrou P,
      3. Lambadiari V, et al
      . Insulin effects in muscle and adipose tissue. Diabetes Res Clin Pract 2011;93 Suppl 1:S52–9. doi:10.1016/S0168-8227(11)70014-6
      OpenUrl

    Supplementary materials

    Footnotes

    • RS and NHA-s are joint first authors.

    • CL-Q and UP-B are joint senior authors.

    • Contributors RS, CL-Q and UP-B contributed to the conceptual design and data analysis of the post hoc study. A-SS contributed to the conceptual design, recruited participants and conducted the original clinical trial. IMM performed the metabolomic analysis. JP contributed to the data analysis. NHA-s and RS contributed equally to the drafting of the manuscript. All authors contributed significantly to the manuscript review and editing and approved the final manuscript. UP-B is the guarantor and is responsible for the overall content of the manuscript.

    • Funding The study was funded by the Novo Nordisk Foundation (grant number NNF17OC0028358). The original trial was funded by research grants from the University of Southern Denmark, the Danish PhD School of Endocrinology, the Research Foundation at Nordsjællands Hospital, HypoSafe, the Fog Foundation, the Augustinus Foundation, the Tvergaard Foundation, the Olga Bryde Nielsen Foundation and the Helen Rudes Foundation.

    • Competing interests After conclusion of the experimental work, A-SS has been employed by Novo Nordisk. CL-q is a member of the Scientific Advisory Board of Fondation Alzheimer and Institute Pasteur, and UP-B has served on advisory panels for Novo Nordisk and Sanofi.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.