Figure 1. Urinary citrate levels from late-onset patients with glycogen storage disease type II. A, Urine levels from 19 untreated patients with glycogen storage disease type II compared with controls (reference upper limit, 803 mmol/mol creatinine). B, Urine citrate at baseline and 7 months later in 4 untreated patients illustrating the intermittent nature of excessive urine citrate levels.
Figure 2. Plasma insulin-like growth factor type 1 (IGF-1) (A) and insulin-like growth factor binding protein 3 (IGFBP-3) (B) levels from 26 untreated patients with glycogen storage disease type II (gray bars) compared with 164 healthy participants (open bars) distributed by age ranges. The levels for both IGF-1 and IGFBP-3 were significantly elevated for the late-onset patients with glycogen storage disease type II at all age ranges indicating potentially impaired intracellular transfer of IGF-1. To convert IGF-1 to nanomoles per liter, multiply by 0.131.
Figure 3. Acute response to a triheptanoin meal over 6 hours by insulin-like growth factor type 1 (IGF-1) (A) and urine citrate (B) levels. The decrease of IGF-1 suggests increased intracellular uptake, possibly related to increased adenosine triphospate availability. The rapid decline of urinary citrate is consistent with enhanced metabolism of citrate, potentially due to enhanced adenosine triphospate production and inactivation of adenosine monophosphate–activated protein kinase. To convert IGF-1 to nanomoles per liter, multiply by 0.131.
Figure 4. Schematic of the integrated methylation pathways. Blue font represents metabolites that accumulate and red font denotes those whose concentration decreases as observed with patients with glycogen storage disease type II. ADP indicates adenosine diphosphate; AGAT, L-arginine:glycine amidinotransferase; ATP, adenosine triphosphate; CK, creatinine kinase; GAA, guanidinoacetate; GAMT, guanidinoacetate N-methyltransferase; MS, methionine synthase; MTHF, methyltetrahydrofolate; SAH, S -adenosylhomocysteine; SAM, S -adenosylmethionine; and THF, tetrahydrofolate.
Pascual JM, Roe CR. Systemic Metabolic Abnormalities in Adult-onset Acid Maltase DeficiencyBeyond Muscle Glycogen Accumulation. JAMA Neurol. 2013;70(6):756-763. doi:10.1001/jamaneurol.2013.1507
Author Affiliations: Rare Brain Disorders Clinic and Laboratory, Departments of Neurology and Neurotherapeutics (Drs Pascual and Roe), Physiology and Pediatrics (Dr Pascual), The University of Texas Southwestern Medical Center, Dallas.
Importance The physiological relevance of acid maltase (acid α-glucosidase, an enzyme that degrades lysosomal glycogen) is well recognized in liver and muscle. In late (adult)–onset acid maltase deficiency (glycogen storage disease type II [GSD II]), glycogen accumulates inside muscular lysosomes in the context of reduced enzymatic activity present not only in muscle, but also throughout the organism. Yet, disease manifestations are commonly attributed to lysosomal disruption and autophagic vesicle buildup inside the myofiber due to a lack of obvious hepatic or broader metabolic dysfunction. However, current therapies primarily focused on reducing glycogen deposition by dietary or enzyme replacement have not been consistently beneficial, providing the motivation for a better understanding of disease mechanisms.
Objective To provide a systematic overview of metabolism and methylation capacity using widely available analytical methods by evaluating secondary compromise of (1) the citric acid cycle, (2) methylation capacity, and (3) nutrient sensor interaction in as many as 33 patients with GSD II (ie, not all patients were available for all assessments) treated with only a low-carbohydrate/high-protein, calorie-balanced diet.
Design, Setting, and Patients Case series including clinical and analytical characterization in an academic setting involving 33 enzymatically proved adults with GSD II treated only with a low-carbohydrate/high-protein, calorie-balanced diet.
Main Outcome and Measure Biochemical analysis of blood and urine samples.
Results Patients exhibited evidence for disturbed energy metabolism contributing to a chronic catabolic state and those who were studied further also displayed diminished plasma methylation capacity and elevated levels of insulin-like growth factor type 1 and its carrier protein insulin-like growth factor binding protein 3 (IGFBP-3).
Conclusions and Relevance The simplest unifying interpretation of these abnormalities is nutrient sensor disturbance with secondary energy failure leading to a chronic catabolic state. Data also provide the framework for the investigation of potentially beneficial interventions, including methylation supplementation, as adjuncts specifically targeted to ameliorate the systemic metabolic abnormalities of this disorder.
Glycogen metabolism occurs throughout the organism, particularly in organs that expend energy generating work and maintaining metabolic homeostasis, such as muscle and liver.1,2 Unlike other glycogenoses, late-onset glycogen storage disease type II (adult GSD II) is characterized by loss of function in all tissues of the lysosomal enzyme, acid α-glucosidase (acid maltase), which tends to exhibit higher residual activity than in the infantile form of the disorder.3 Despite this uniform deficiency, patients with GSD II do not experience multiorgan failure. Instead, striated skeletal and sphincter muscles (especially the diaphragm) are severely affected, causing dysphagia, sphincter hypotonia, progressive limb weakness, muscular atrophy, and ultimately, respiratory failure.4,5
Elevations of serum enzymes (creatine kinase [CK] and transaminases) and, occasionally, decreases in plasma alanine and glutamine levels, have generally been attributed to the myopathy,6,7 such that there has been no systematic focus on the potential contribution of additional organ derangement to the disorder.8 Consequently, current management is centered on limiting glycogen deposition in skeletal muscle via dietary modification or enzyme replacement (ERT). The current diet includes reduced carbohydrate and increased protein intake, often with alanine supplementation and programmed physical exercise.6,7,9- 12 This diet, combined with exercise, reversed muscle glycogen accumulation in 2 affected siblings11 and has retarded the rate of clinical deterioration.10 Following introduction of intravenous ERT, clinical studies have not demonstrated homogeneous benefit for patients with GSD II.13,14 The obvious corollary to these observations is that therapy for GSD II may be insufficient to provide the desired clinical benefit. From a more critical perspective, they may also be taken to suggest the existence of additional, unrecognized, and uncorrected metabolic derangements (other than compromised glycogen degradation) that may contribute to the pathogenesis of this disorder and, perhaps, to an inadequate response to current therapies.13
In this study, we set out to provide a systematic overview of metabolism and methylation capacity using widely available analytical methods by evaluating secondary compromise of (1) the citric acid cycle (CAC), (2) methylation capacity, and (3) nutrient sensor interaction in as many as 33 patients (ie, not all patients were available for all assessments) treated only with diet. We reasoned that, if affected, these factors may reflect an underlying energy-deficient state contributing to the progressive clinical deterioration of these patients.
Standard informed consent and ethical procedures were followed. One patient received triheptanoin supplementation solely for analytical purposes (rather than to test clinical efficacy) under an instutional review board–approved protocol and Food and Drug Administration IND 59303. Lysosomal α-glucosidase deficiency was confirmed in all patients by leukocyte enzymatic assay prompted by myopathic clinical features. All available patients known to us with disease onset after age 15 years were included. They were sequentially enrolled during 1 year on the basis of clinical and enzymatic criteria. Not all patients were available for participation in all analytical studies, and additional selection criteria were not imposed on those who were able to participate in further study. Disease onset was defined as the time when weakness or fatigability first led to general medical attention. Additional data included demographic profiles, ambulatory and ventilatory assistance, nutritional route, and Walton and Gardner-Medwin14 and Slonim10 functional scores.
All blood assays were performed on plasma obtained from (overnight) fasting samples. Methods for quantitative acylcarnitines, plasma amino acids, and urinary organic acids, and their reference ranges have been described previously.15- 17 Acylcarnitine profiles and plasma amino acid levels were obtained in all 33 patients. Samples for urine organic acid analyses were available in a subset of 19 patients. Blood chemistry analyses included the following levels: glucose, serum urea nitrogen, creatinine, aspartate aminotransferase, alanine aminotransferase, and CK.
Total plasma homocysteine was measured by high-performance liquid chromatography with fluorescence detection.18 Plasma S -adenosylmethionine (SAM), and S -adenosylhomocysteine (SAH) were measured in 3 patients by a modification of the stable-isotope dilution liquid chromatography–electrospray injection tandem mass spectrometry previously described.19 Plasma guanidinoacetic acid, and creatine were measured in 7 patients by the Clinical Chemistry Department of the Free University of Amsterdam, Amsterdam, the Netherlands, using previously described methods.20
Plasma IGF-1, insulin-like growth factor binding protein-3 (IGFBP-3), and growth hormone (GH) were measured in all patients by D. S. Laboratories, Houston, Texas. The results were compared to age-related standards because both IGF-1 and IGFBP-3 decrease with age.
Statistical results were obtained with GraphPad Prism (version 4.00; GraphPad Software, Inc) and tested using a nonparametric Mann-Whitney test. Results were considered statistically significant at P ≤ .05.
Patient age ranged from 15 to 72 years and included 15 females and 18 males whose symptom duration ranged from 2 to 48 years (Table 1). They were sequentially enrolled during 1 year on the basis of clinical and enzymatic criteria. As noted, not all patients were available for participation in all analytical studies, and the tables allow for correlative patient identification. Of the 33, at the time of blood and urine testing, 10 required a wheelchair, 18 were ambulatory, and 5 were ambulatory but required assistive devices. Ten patients (30%) required respiratory support (8 nightly, 2 continuously). Frozen samples for metabolic analysis (propionylcarnitine, alanine, and glutamine) were obtained from all 33 patients. Urinary citrate was measured in 26 patients. All patients received a standard low-carbohydrate/high-protein, calorie-balanced diet.6 None of the patients had received ERT or supplements, such as carnitine, alanine, glutamine, or citrate. Although follow-up analyses were beyond the scope of this study, 4 patients were tested again after 7 months, to evaluate whether urinary citrate excretion was subject to fluctuation.
Quantitative results for blood propionylcarnitine, alanine, glutamine, and urinary excretion of citrate are summarized in Table 2 together with patient ages. No significant or consistent abnormalities were noted in the plasma amino acid or urinary organic acid analyses. Quantitative blood spot free carnitine and acylcarnitines (acetyl-[C2]) to linoleoyl-carnitine (C18:2) were measured for all patients. Propionylcarnitine was markedly reduced in most patients. Using our isotope dilution method, levels less than 1.5 μM and, especially, less than 1.0 μM are rarely observed in healthy participants in blood spot analyses (C.R.R., unpublished data, August 18, 2012). Twenty-three (70%) of 33 patients exhibited levels lower than 1.5 μM (reference range, 0.28-1.42 μM) and 11 patients exhibited levels lower than 1.0 μM (reference range, 0.28-0.95 μM). All other acylcarnitine levels were consistently normal (data not shown). Plasma amino acid analysis revealed that only 3 patients exhibited reduced levels of alanine or glutamine. Serum CK levels for all 33 patients ranged from 58 to 1401 U/L (to convert units per liter to microkatals per liter, multiply by 0.0167). Twenty-four patients (73%) exhibited elevated CK levels (>130 U/L).
Urine samples for organic acid analysis were available from 26 patients. Only citrate levels demonstrated abnormalities. Citrate levels from 12 patients (46%) were above normal. Eight of these were markedly increased (>1000 mmol/mol creatinine) (Figure 1A). Due to the wide distribution of citrate levels in these patients, the group, as a whole, was not significantly different from healthy controls (P = .09). However, comparison of the 8 patients with citrate levels higher than 1000 mmol/mol creatinine revealed a significant difference from healthy controls (P ≤ .001). Four of 8 patients were studied again 7 months later. Initially, 3 of these 4 patients had urinary citrate levels in the reference range (<803 mmol/mol creatinine). However, 7 months later, their urine citrate levels had become elevated (Figure 1B), indicating that citrate excretion was subject to considerable fluctuation. No other organic acid abnormalities were noted to suggest metabolic derangement in fat oxidation, amino acid degradation, or vitamin deficiencies.
Plasma levels of IGF-1, IGFBP-3, and GH were measured in 26 patients who were available for further analytical testing (Figure 2). Plasma levels of IGF-1 and IGFBP-3 gradually decrease with age. Therefore, patient results were compared with 3 healthy age groups (30-40 years, 41-50 years, and 51-70 years). In comparison with the corresponding reference ranges for age, all 26 patients exhibited significant elevations in both IGF-1 and IGFBP-3. The IGF-1 levels were highly significant for all age groups (P ≤ .001). The comparison of IGFBP-3 levels with normal age ranges was also significant: P = .02 for the 30- to 40-year-old patients, and P ≤ .001 for the 2 older age groups. Growth hormone levels were normal (<0.1 ng/mL) (to convert nanograms per milliliter to nanomoles per liter, multiply by 0.131) for all 26 patients (data not shown).
The previous data led to the hypothesis that, if IGF-1 and IGFBP-3 levels were elevated as the result of energy deficiency in GSD II, stimulation of anaplerosis (ie, replenishment of CAC intermediates) might reduce both plasma IGF-1 and IGFBP-3 levels and reduce urinary citrate excretion. To test this hypothesis, the effects of a single oral triheptanoin meal (an anaplerotic triglyceride21- 23) were tested in 1 patient. This 65-year-old man was given a single dose of triheptanoin (0.25 g/kg) in 80 g of low-fat/low-carbohydrate yogurt. Plasma IGF-1, IGFBP-3, GH, and urinary citrate were measured over 6 hours following the meal. The IGF-1 levels decreased by 23%, from 213 ng/mL to a normal mean level of 163 ng/mL (IGF-1 reference, 154[ 49] ng/mL) (Figure 3A). The IGFBP-3 levels also decreased to a normal level for age: from 3672 to 3345 ng/mL. Growth hormone levels remained normal at less than 0.1 ng/mL. Urinary citrate excretion also progressively decreased by 19% from 463 to 373 mmol/mol creatinine (Figure 3B).
The integrity of methylation was assessed in a subset of 7 patients (Table 3). In 3 of these patients, plasma levels of total homocysteine, SAM, and SAH, and the SAM/SAH ratio were assayed. Homocysteine levels were normal. The levels of SAM were decreased, and the SAH levels were elevated, resulting in a reduced ratio of SAM/SAH. All 7 patients had plasma levels of creatinine, methionine, guanidinoacetate, and creatine additionally determined. Plasma methionine levels were variable ranging from 12 to 47 mM (reference range, 21-35 mM). Guanidinoacetate levels were normal in all patients. However, all plasma creatinine levels were reduced below normal values, while plasma creatine levels were markedly elevated in all 7 patients. These elevations in creatine levels are too marked to be simply the consequence of increased protein dietary intake.24
Other than simply muscle glycogen deposition and vesicular myofiber disruption, lines of evidence for a more extensive and complex metabolic dysfunction in GSD II may have, in retrospect, been present previously and are expanded in this study. Together, these findings suggest an underlying energy deficiency producing a chronic catabolic state with the potential to significantly impact skeletal muscle function and preservation. These lines of evidence are discussed sequentially.
In GSD II, striated muscle wasting compromises physical performance and leads, ultimately, to respiratory failure. Unlike the infantile form, the late-onset disease does not usually affect the liver or heart. Potential hepatic abnormalities in GSD II have been limited to mild increases in serum transaminase levels that, in fact, may be due to myopathy. Other liver dysfunction indicators, such as hypoglycemia and hyperammonemia, are not features of GSD II.3 Conventional low-carbohydrate/high-protein, calorie-balanced dietary therapy alone has not been uniformly successful. Encouraging results have been reported with enhanced protein/low-carbohydrate intake associated with a consistent physical therapy program in the few patients studied.6,10 Enzyme replacement has shown benefits in mobility and stability of the clinical course in the first year but without reversal of respiratory compromise or continued benefits during the second year. Unfortunately, ERT has not been consistently beneficial in all patients with GSD II.25 The limited benefits of these treatments suggest that the metabolic consequences of this disorder are more extensive than previously recognized and require further evaluation to improve management in a broader metabolic context.
Residual enzyme activity in postmortem studies is significantly reduced in all organs tested.26 Despite the fact that the enzyme deficiency is severe in both skeletal muscle and liver, glycogen deposition is almost confined to skeletal muscle with little, if any, in liver.8 A hypothesis compatible with these findings is that the substrate requirements for the hepatic CAC and associated energy production may be provided at the expense of skeletal muscle protein degradation, contributing to consequent deterioration. Nevertheless, these observations also support the need to develop additional treatment strategies focused on amelioration of secondary but closely interrelated biochemical abnormalities.
Creatinine kinase was increased in 24 (73%) of the 33 patients along with occasional increases in serum transaminase levels. Plasma creatinine levels were below the reference range in all 7 patients studied. Blood acylcarnitine analyses revealed reduced levels of propionylcarnitine in 23 of 33 patients (<1.50 μM). No other evidence was noted for disturbed propionate metabolism. Reduced blood levels of propionylcarnitine can reflect overconsumption of propionyl-coenzyme A (CoA) to augment succinyl-CoA in the CAC in an energy-deficiency state. This decrease to less than 1.50 μM is often observed in patients with other inherited disorders, such as pyruvate carboxylase, adult-onset carnitine palmitoyltransferase II, and glycogen brancher deficiencies.22,23
Urinary citrate was the only CAC intermediate that was increased in 8 of 19 patients (Table 2). Urinary citrate levels were higher than 1000 mmol/mol creatinine in these patients (Figure 1A), exceeding normal urinary citrate levels (typically <803 mmol/mol creatinine) (C.R.R., personal observations, August 18, 2012). Despite this finding, no significant difference was noted (P = .09) from healthy controls. However, 4 of 19 patients had additional levels assayed 7 months later revealing that urinary citrate levels can significantly vary over time and are not a persistent abnormality for individual patients. Three of these 4 patients exhibited normal citrate levels when first analyzed, which increased significantly 7 months later, possibly reflecting changes in their metabolic state or disease progression (Figure 1B). When mitochondrial citrate enters the cytosol, it must first be converted by adenosine triphosphate (ATP)–citrate lyase (lyase) to acetyl-CoA and oxaloacetate. Acetyl-CoA produced by the lyase reaction can be converted by either acetyl-CoA carboxylase II (ACC II) to produce malonyl-CoA (inhibiting β-oxidation) or be used by acetyl-CoA carboxylase I (ACC I) to facilitate fatty acid synthesis.27,28 Impairment of these reactions due to reduced ATP availability would lead to urinary citrate excretion. Of note, and related, adenosine monophosphate (AMP)–activated protein kinase (AMPK) activation by reduced ATP stimulates cytosolic catabolic pathways that normally enhance ATP synthesis while inhibiting biosynthetic reactions that consume ATP. The result is a catabolic state. Active AMPK inhibits both ACC I and ACC II, and may contribute to the intermittent excessive urinary excretion of cytosolic citrate observed in patients with GSD II.
As observed in the adult form of GSD IV (adult polyglucosan body disease),22 there is also evidence for a secondary impairment of the integrated pathways of methylation in GSD II (Table 3). These abnormalities included a reduced SAM level, increased SAH level, elevated creatine level, and reduced creatinine level. These integrated pathways require ATP as depicted in Figure 4. The increased plasma levels of creatine associated with decreased levels of creatinine suggest compromised synthesis and availability of creatine phosphate that also requires ATP. This observation further supports the existence of a potential compromise of intracellular energy metabolism. Similar abnormalities were observed in patients with GSD IV in whom, additionally (and except for creatinine levels), normalization was observed following 6 months of the anaplerotic triheptanoin diet.22
In late-onset GSD II lysosomal acid α-glucosidase is the only defective enzyme involved in glycogen degradation. Although designated as a glucosidase, the enzyme also displays acid-debrancher activity.29 Therefore, it can catalyze the complete hydrolysis, at acid pH, of its natural substrate, glycogen. However, because the cytoplasmic glycogenolytic enzymes (active at neutral pH) are unaffected in GSD II, the progressive glycogen deposition in the cytosol as well as in lysosomes remains unexplained. The cytosolic glycogen accumulation might result from continued glycogen synthesis due to ineffective regulation of glycogen synthase by several ATP-dependent protein kinases, including AMPK.30- 32 This could lead to continued glycogen deposition in the cytosolic compartment in excess of its degradation by the intact cytosolic glycogenolytic enzymes. This possibility has also been suggested for late-onset GSD IV22 and McArdle disease (GSD type V).33
Healthy individuals receiving high-protein diets exhibit modest increases in IGF-1 levels. For them, each 1-SD increment in total protein, dairy protein, and calcium intake is associated with an increase in plasma IGF-1 levels of only approximately 2.5%. However, IGFBP-3 levels are unaffected.34,35 Plasma IGF-1 levels in patients with GSD II were much greater at more than 84% above normal mean levels for all age groups (P < .0001). Therefore, the significantly elevated plasma levels of both IGF-1 and IGFBP-3 in patients with GSD II in the context of normal growth hormone levels are not due to high-protein intake.
These extreme plasma levels of IGF-1 in GSD II may also reflect a disturbance in nutrient sensor interactions based on energy deficiency in muscle metabolism. The most abundant IGF-1 binding protein is IGFBP-3, which binds 90% of all circulating IGF-1.34 The remarkable plasma IGF-1 elevations are compatible with dysfunctional entry of IGF-1 into muscle cells via the ATP-requiring tyrosine kinase–IGF-1 receptor. Normally, intracellular IGF-1 inhibits the catabolic effects of active AMPK via serine-threonine kinase and permits activation of the mammalian target of rapamycin. This stimulates biosynthetic reactions, cell proliferation, DNA synthesis, uptake of amino acids and glucose, and suppression of proteolysis.27,36 Consistent with these observations, mice lacking the functional IGF-1 receptor are small (45% of normal body weight) and die soon after birth of respiratory failure.37 Functional impairment of this receptor could lead to proteolysis and progressive deterioration of muscle resulting in muscle wasting and respiratory insufficiency as manifested by patients with GSD II.
Because the metabolic derangements in GSD II suggest an energy-deficient catabolic state, the anaplerotic triglyceride triheptanoin at 35% of total caloric intake has been tested. The significant effects in a 42-year-old woman of protein sparing, decreased proteolysis, and reversal of acute respiratory failure coupled with a return to a normal life style have been previously described by us.38 However, IGF-1 levels were not measured. In the current study, a 65-year-old man received a single dose of triheptanoin (0.25 g/kg) and was evaluated over 6 hours in an attempt to observe any effect of anaplerosis on plasma IGF-1 and IGFBP-3 levels and urinary citrate excretion. His plasma IGF-1 and IGFBP-3 levels decreased to normal levels. During the same period, his urinary citrate excretion also decreased (Figure 3). These effects may reflect the consequence of enhanced anaplerosis on the IGF-1 receptor and normalization of citrate metabolism, which are both dependent on enhanced ATP availability. In the context of all the results described earlier, this single observation and the related article38 provide justification for further, systematic evaluation of anaplerotic therapy in this disorder.
In conclusion, these observations in patients with adult-onset GSD II suggest that there is a significant energy deficit in this disease that is reflected in metabolic abnormalities including reduced methylation capacity. Thus, as also suspected by others, the pathogenetic mechanisms of GSD II seem to be broader than they were generally thought to be.13 The most economic interpretation of our results is that these abnormalities may be related to compromised regulation of nutrient sensors, producing a chronic intermittent catabolic state. These observations, together with our prior findings involving triheptanoin effects in GSD II and late-onset GSD IV22,38 suggest that anaplerotic diet therapy and methylation supplements may assist in the future management of patients with GSD II.
Correspondence: Juan M. Pascual, MD, PhD, Rare Brain Disorders Clinic and Laboratory, The University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Mail Code 8813, Dallas, TX 75390.
Accepted for Publication: October 23, 2012.
Published Online: April 22, 2013. doi:10.1001/jamaneurol.2013.1507
Author Contributions:Acquisition of data: Roe. Analysis and interpretation of data: All authors. Drafting of the manuscript: All authors. Critical revision of the manuscript for important intellectual content: All authors. Statistical analysis: Roe. Administrative, technical, and material support: Roe. Study supervision: Roe.
Conflict of Interest Disclosures: None reported.
Funding/Support: The study was supported in part by the Dallas Women's Foundation (Billingsley Fund) (Dr Pascual), and the Baylor Health Care System Foundation from generous donations of Mr William Hutchison (Dr Roe).
Additional Contributions: Arnold Reuser, PhD, and Ans van der Ploeg, MD, PhD, Department of Clinical Genetics, Erasmus Universiteit, Rotterdam, the Netherlands, provided untreated samples from 26 patients for this study. Teodoro Bottiglieri, PhD, Institute of Metabolic Disease, determined SAM, SAH, methionine, homocysteine, IGF-1, IGFBP-3, and GH plasma levels. Cornelis Jakobs, PhD, Clinical Chemistry Department of the Free University of Amsterdam, Amsterdam, the Netherlands, analyzed the guanidinoacetate and creatine levels.
Additional Information: Part of this study was performed at the Institute of Metabolic Disease, Baylor University Medical Center, Dallas.