| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
From the Clinical Research Centers |
Divisions of Medical Genetics, Department of Pediatrics, Emory University, Atlanta, Georgia 30322
Address all correspondence and requests for reprints to: Nicola Longo, M.D., Ph.D., Division of Medical Genetics, Department of Pediatrics, Emory University, 2040 Ridgewood Drive, Atlanta, Georgia 30322.
 |
Abstract |
|---|
 Top Abstract IntroductionExperimental SubjectsMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionExperimental SubjectsMaterials and MethodsResultsDiscussionReferences |
|---|
-subunits and two
ß-subunits, linked by disulfide bonds (1, 2). Insulin binding to the
-subunit activates ß-subunit autophosphorylation and kinase
activity. This kinase activity is essential for transmembrane signaling
of glucose transport and for activating intracellular substrates of
insulin action, such as insulin-responsive substrate-1 (IRS-1), IRS-2,
IRS-3, and IRS-4 (3). Both subunits of the insulin receptor are encoded
by a single gene located on the short arm of chromosome 19 (4).
Mutations in this gene cause inherited insulin-resistant syndromes
whose severity ranges from mild to severe. Extreme insulin resistance is observed in patients with leprechaunism and Rabson-Mendenhall syndrome (5, 6, 7). These are autosomal recessive conditions in which both alleles for the insulin receptor are abnormal, and patients fail to respond to endogenous and exogenous insulin. Affected patients have intrauterine and postnatal growth restriction (resulting from the defective mitogenic action of insulin), dysmorphic features, lack of sc fat, acanthosis nigricans, enlargement of genitalia, hirsutism, paradoxical fasting hypoglycemia and postprandial hyperglycemia, and extremely elevated levels of circulating insulin, up to 1000 times above normal. Children with leprechaunism usually die before 1 yr of age. They do not present diabetic ketoacidosis, and although they have post-prandial hyperglycemia, their major metabolic problem is fasting hypoglycemia. Rabson-Mendenhall syndrome differs from leprechaunism in the presence of premature and dysplastic dentition (which is sometimes observed at birth), coarse facial features, and pineal hyperplasia (which is of unknown significance) (5, 6, 7). In addition, children with Rabson-Mendenhall syndrome develop with time severe and intractable diabetic ketoacidosis. The reason for the paradoxical hypoglycemia and the development of ketoacidosis is unclear.
Studies performed in patients with leprechaunism suggested that fasting hypoglycemia was caused by a combination of insulin resistance and an accelerated fasting state with rapid depletion of hepatic glycogen (8). However, subsequent ultrastructural studies of the liver indicated the presence of normal to increased amounts of glycogen during fasting (9), which was more consistent with defective hepatic glucose release.
Recently, animal models for these conditions were obtained by homologous recombination and knock out of the insulin receptor gene (10, 11). Unfortunately, these animals die shortly after birth of diabetic ketoacidosis and do not present the paradoxical fasting hypoglycemia, shedding little light on this phenomenon.
In a previous study we reported a patient with Rabson-Mendenhall syndrome whose growth and metabolic parameters failed to respond to GH and insulin-like growth factor I (IGF-I) (12). In this report, we describe the progression of this patient to ketoacidosis. Our results indicate that this progression is associated with a decline in plasma insulin levels. Extremely elevated insulin levels were measured only early in life and were associated with fasting hypoglycemia, suggesting that the excessive amounts of insulin during fasting may represent the basis for hypoglycemia in these patients with abnormal insulin receptors.
|
Experimental Subjects |
|---|
|
TopAbstractIntroductionExperimental SubjectsMaterials and MethodsResultsDiscussionReferences |
|---|
At about 7 yr of age, the child presented with fever and irritability. He was given parenteral antibiotics, and while transport to a tertiary center was being arranged, the patient developed intractable seizures. Upon arrival at the hospital, a pentobarbital coma was induced to control seizures. Laboratory evaluation indicated a glucose level above 700 mg/dL and compensated ketoacidosis. A lumbar puncture performed the next day (for the extreme instability) was consistent with meningitis, although routine cerebrospinal fluid cultures were negative. Urine analysis and urinary organic acids were positive for massive amounts of ketones. To control the hyperglycemia and ketonuria, insulin infusion was started and increased up to 9.5 U/kg·h, when the plasma insulin level reached 10,173 µU/mL. At this dose, plasma glucose levels remained mildly elevated (200–400 mg/dL), and ketonuria abated. The child had a protracted course in the intensive care unit and after 3 weeks expired of sudden cardiac arrest. Autopsy was not performed.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionExperimental SubjectsMaterials and MethodsResultsDiscussionReferences |
|---|
Data were collected from chart review. Glucose and insulin were measured by standard procedures performed by reference laboratories. After plotting, data were analyzed by linear or nonlinear regression using Excel. Significance of regression analysis was established using ANOVA.
Analytical methods
Urinary organic acids were analyzed by gas chromatography/mass spectroscopy according to standard methods (13). Urine from patients with type 1 diabetes (aged 6–8 yr) was obtained during acute hyperglycemia.
Identification of mutations in the insulin receptor gene was performed by direct sequencing each of the 22 exons of the insulin receptor, as previously described (14, 15). Mutations were confirmed by restriction analysis of PCR-amplified genomic DNA.
|
Results |
|---|
|
TopAbstractIntroductionExperimental SubjectsMaterials and MethodsResultsDiscussionReferences |
|---|
Sequencing of the insulin receptor gene revealed that patient
Atl-2 was a compound heterozygote for two mutations in exon 19. The
maternal mutation was a T to C transition converting the codon for
Ile1131 (ATT) to Thr (ACT; I1115T). This
mutation created a novel AlwNI restriction site in exon 19
(CTGNNN GAC), which was cleaved into two fragments
of 103 and 182 bp in one of the proband’s and one of the mother’s
alleles (Fig. 1
). The paternal mutation
was a C to T transition converting the codon for Arg1131
(CGG) to Trp (TGG; R1131W). This mutation abolished a
SfaNI restriction site (GCATCNNN ) that
normally cleaves exon 19 into two fragments of 142 and 143 bp, visible
as a single band in Fig. 1. The father and the proband were
heterozygotes for this mutation and presented an additional uncleaved
285-bp band. No other variations from the published sequence and known
polymorphisms were observed in the remaining 21 exons.
|
Patient Atl-2 presented with acute ketoacidosis for the first time
at 6 yr of age. Urinary organic acids indicated the presence of
excessive compounds from ß-oxidation, with high levels of
acetoacetate and 3-hydroxybutyric acid (Table 1). The accumulation of dicarboxylic
acids and 3-hydroxy intermediates (see list in Table 1) reflects the
increased lipolysis and microsomal 
-oxidation (16). Urine from
children with type 1 diabetes exhibited alterations similar to those of
the patient with Rabson-Mendenhall syndrome, although to a lesser
degree.
|
The similarity of the organic acid profile suggested that the patient with Rabson-Mendenhall syndrome had insulin deficiency (relative compared to absolute), as in patients with type 1 diabetes. Therefore, plasma insulin levels were measured and were still elevated (>500 µU/mL; 3.5 nmol/L) at the time of ketoacidosis. To gain further understanding of the mechanism producing ketoacidosis in this patient, plasma insulin and glucose levels were reviewed since birth.
At birth (Fig. 2), there was both fasting
hypoglycemia (glucose, 1.1 mmol/L; 20 mg/dL) and postprandial
hyperglycemia (glucose, 13.9 mmol/L; 250 mg/dL) with very high levels
of circulating insulin (>8000 µU/mL; 57.4 nmol/L; normal for the
reference laboratory, up to 21 µU/mL; 150 pmol/L). Therapy at that
time was directed at preventing hypoglycemia with frequent feedings and
a diet rich in complex carbohydrates. At about 4 yr of age,
hypoglycemia was no longer present, and only hyperglycemia was
observed. No ketosis could be seen at this time even with urinary
organic acid analysis. Insulin levels were decreased, on the average,
compared to those at birth (from 41.0 ± 9.4 to 10.4 ± 1.2
nmol/L; P < 0.05 using 95% confidence intervals). At
6 yr of age, when he presented ketoacidosis, insulin levels remained
elevated, but were further decreased since birth (3.19 ± 0.85
nmol/L; P < 0.05 vs. values at birth and at
4 yr of age using 95% confidence intervals).
|
) indicated that insulin levels
significantly (P < 0.001) decreased with age. At the
same time (Fig. 2B) there was a significant (P <
0.005) increase in plasma glucose levels with aging. Despite the
statistical significance, only portion of the variability was explained
by the regressions (r2 = 0.52 for insulin and
r2 = 0.25 for glucose), suggesting that aging was not
the sole factor explaining these changes.
Glucose is the major physiological stimulus for insulin production and
secretion by the pancreatic ß-cell. Therefore, insulin levels were
plotted as a function of blood glucose concentrations during the
lifetime of our patient. When all values were considered, no
significant correlation (P = 0.91; r2
< 0.01; not shown) could be found. However, the patient had
significant clinical changes during his lifetime, and the relationship
between glucose and insulin levels could have changed during the
different pathophysiological states. To test this hypothesis, insulin
levels were plotted as a function of plasma glucose concentrations
during the period of hypo- and hyperglycemia (birth to 1 yr),
hyperglycemia without hypoglycemia (3–4 yr), and hyperglycemia with
ketoacidosis (after 6 yr). As shown in Fig. 3, the regression was statistically
significant (P < 0.05) at 0–1 yr of age (Fig. 3A) and
at 6–7 yr of age (Fig. 3C) and was highly significant
(P < 0.01) between 3–4 yr of age (Fig. 3B). Between
birth and 1 yr of age, insulin increased steeply [the slope of the
regression line (b), 2.32] with increasing glucose concentration, and
at a predicted glucose value of 0, the insulin concentration was still
significantly elevated (
20 nmol/L; Fig. 3A). When only hyperglycemia
was present (3–4 yr of age), insulin increased less steeply (b, 0.74),
and the intercept on the y-axis became very close to 0 (Fig. 3B). A further decrease in the slope (b, 0.16) was observed when the
child became ketoacidotic. These results indicate that insulin levels
reflected plasma glucose concentrations, but that the relationship
between glucose and insulin changed over time, with a progressive
decrease in insulin levels at the same glucose concentration.
|
),
hormone levels were normalized for plasma glucose and plotted as a
function of age (Fig. 4). The
insulin/glucose ratio decreased exponentially with age
(P < 0.0001; highest level of significance tested),
and almost all of its variability was explained by the regression
(r2 = 0.92), indicating that the major confounding
factor in Fig. 2 was the variation in insulin with different glucose
levels. There was first a rapid fall of the insulin/glucose ratio up to
about 2 yr of age, reducing the insulin/glucose ratio to less than 20%
of that measured shortly after birth. After this rapid fall,
hypoglycemia disappeared. After 4 yr of age, there was a slow, more
progressive decrease in insulin levels, with no clear demarcation
indicating the period during which ketoacidosis became evident.
|
|
Discussion |
|---|
|
TopAbstractIntroductionExperimental SubjectsMaterials and MethodsResultsDiscussionReferences |
|---|
Patient Atl-2 had classical clinical and dysmorphological features of
patients with Rabson-Mendenhall syndrome (12). Analysis of the insulin
receptor gene confirmed compound heterozygosity for two mutations
affecting the ß-subunit of the insulin receptor: I1115T and R1131W.
These mutations result in partially reduced insulin binding to the
patient’s cells, which remains about 25% of the levels in control
cells (12). Although none of these mutations has been expressed in
heterologous cells, both affect conserved regions of the tyrosine
kinase domain of the insulin receptor. The I1115T mutation affects a
hydrophobic amino acid that is conserved in the IGF-I, epidermal growth
factor, c-Src, and Ret tyrosine kinases (17, 18). The fibroblast growth
factor receptor 1, Met, Kit, platelet-derived growth factor-,
FMS-related tyrosine kinase 1, and Trk kinases have a different
hydrophobic residue (valine) (17, 18). Interestingly, substitution of
the valine residue in this position of Met with a different hydrophobic
amino acid, leucine, is found in hereditary forms of papillary renal
carcinoma and results in constitutive activation of the Met kinase
(18). The mutated Met complementary DNA does not confer focus-forming
ability to NIH-3T3 cells, but renders them tumorigenic when injected
into nude mice (18). These results stress the importance of hydrophobic
residues in this portion of tyrosine kinase receptors and suggest that
substitution with a hydrophilic amino acid, such as threonine, may
impair insulin receptor kinase activity. The R1131W substitution
affects an amino acid in the core of the catalytic loop of tyrosine
kinase receptors that is conserved in all of the tyrosine kinases
listed above (17). Expression of a less severe substitution of arginine
1331 (R1131Q), which was found in another patient with insulin
resistance, abolished insulin receptor autophosphorylation, kinase
activity toward exogenous substrates, and stimulation of biological
functions (19, 20), suggesting that the less conservative tryptophan
(W) substitution should have even more profound effects on kinase
function. Nonetheless, the preservation of minimal insulin binding by
the patient’s cells is consistent with some residual insulin action
and a phenotype slightly less severe than that of leprechaunism.
Patients with Rabson-Mendenhall syndrome develop ketoacidosis (21, 22, 23, 24). Previous studies have not evaluated the nature of the ketones produced by these patients or the reason for this progression. Our results indicate that urinary organic acids of patients with Rabson-Mendenhall syndrome during ketoacidosis resemble those of patients with diabetes. In addition to excess ketones, a variety of partially oxidized fatty acids are observed, indicating excessive entry of fatty acids inside mitochondria. These results indicate that insulin is unable to suppress the release of fatty acids from adipocytes and that carnitine palmitoyl transferase 1, the rate-limiting step of mitochondrial ß-oxidation, is fully activated in these patients as in patients with diabetic ketoacidosis. This finding is not unexpected, as levels of glucagon, which activates carnitine palmitoyl transferase 1, are normal or slightly increased in patients with Rabson-Mendenhall syndrome (21, 24), including ours (30 and 61 pg/mL; normal range, 50–100 pg/mL).
In contrast to patients with type 1 diabetes, ketoacidosis was observed despite elevated insulin levels in our patient with Rabson-Mendenhall syndrome. However, these insulin levels were much lower than those measured early in life. Early in life, insulin levels were extremely elevated even during periods of paradoxical hypoglycemia. From regression analysis, predicted insulin levels at a glucose level of 0 were 18.2 nmol/L, a value at least 100 times above normal. The fasting hypoglycemia observed at birth disappeared at 3–4 yr of age when the predicted insulin concentration at a glucose level of 0 was not significantly above 0. This result suggests that the paradoxical fasting hypoglycemia observed in children with leprechaunism and early in life in Rabson-Mendenhall syndrome may be due to the extremely elevated levels of circulating insulin. In these children, insulin levels are excessive during fasting when glucose is no longer provided by feedings. Glucose release by the liver is responsible for the maintenance of adequate glucose levels during fasting. This function is normally suppressed by insulin, acting on both the liver and extrahepatic tissues (25). It is possible that this function of liver and extrahepatic tissues remains sensitive to the elevated insulin levels measured in our patient shortly after birth at the time of hypoglycemia. Other functions, such as stimulation of glucose uptake and utilization by peripheral tissues (muscle and fat) at the time of feeding, were not significantly sensitive even to supraphysiological insulin levels, with postprandial hyperglycemia noted immediately or shortly after birth.
As fasting hypoglycemia is also observed in children with no functional insulin receptors, such as patients compound heterozygotes for two null mutations in the insulin receptor (26, 27, 28), the effect of extremely elevated levels of insulin on glucose homeostasis must be mediated by the interaction with receptors different from the insulin receptor, possibly the IGF-I receptor or other as yet uncharacterized receptors.
Fasting hypoglycemia is not observed in mice homozygous for a null allele of the insulin receptor gene (10, 11). The phenotype of these animals differs from that of patients with leprechaunism in the absence of neonatal growth restriction and the presence of hyperglycemia and ketoacidosis at birth (10, 11). Insulin levels in these mice are not elevated to the degree they are in humans with mutant insulin receptors. This is probably due to the relative delay in the appearance of typical ß-cells in the pancreas of mice (0.5 day before the end of gestation) compared to humans, where a significant increase in insulin levels is recorded after 25 weeks gestation (29). Human patients have insulin levels up to 1000 times above normal. By contrast, mice homozygous for the deletion of the insulin receptor gene have insulin levels that are at most 10 times above normal (10, 11). This slight increase in insulin levels is comparable to that seen in our patient after the development of ketoacidosis at 6 yr of age. Therefore, it is conceivable that the neonatal onset of constant hyperglycemia and ketoacidosis in mice homozygous for null alleles for the insulin receptor reflects their relative insulinopenia at birth. In these mice, IGF-I injected immediately after birth decreased glucose levels by about 50% (30). This effect on glucose levels was associated with an IGF-I-mediated decrease in the expression of the key gluconeogenetic enzyme phosphoenolpyruvate carboxykinase (30). As these mice lack any functional insulin receptor, these effects of IGF-I must be mediated through interaction with the IGF-I receptor.
As Rabson-Mendenhall syndrome progressed, insulin levels decreased and were no longer capable of suppressing hepatic glucose production and release, with constant hyperglycemia. When insulin levels decreased further, their capacity to suppress fatty acid oxidation was compromised, and constant ketoacidosis ensued. However, infusion of extremely high concentrations of insulin by continuous infusion (9.5 U/kg·h) reversed increased fatty acid oxidation and blocked ketonuria. These elevated insulin doses may have worked through the residual insulin receptors present in our patient or, more likely, through alternative receptors. Administration of very high doses (1.6 mg/kg·day) of IGF-I improved glucose homeostasis in a patient with extreme insulin resistance due to a deletion of exon 5 and an L87P substitution in the insulin receptor (31). However, only fasting, not postprandial, hyperglycemia is improved by IGF-I, with free fatty acids remaining normal in the IGF-I-treated patient (31). Therefore, a decreased supply of extrahepatic energy may be responsible for the improvement in fasting glucose levels. IGF-I (at more physiological doses) was effective in decreasing free fatty acid levels and the rise in ß-hydroxybutyrate in a patient homozygous for P193L substitution in the insulin receptor (32). As the mutation in this patient abolished insulin binding, IGF-I action was not due to spillover activation of the insulin receptor (32). The doses of IGF-I tested in this latter study did not affect glucose levels, indicating that lipid metabolism is more sensitive to IGF-I action than glucose metabolism (32). Our reversal of ketoacidosis by high doses of insulin in this patient with Rabson-Mendenhall syndrome could therefore be explained by nonspecific activation of IGF-I signaling on lipid metabolism by the elevated concentrations of insulin (73 nmol/L) obtained with iv insulin infusion.
Other mechanisms may contribute to the disappearance of the hypoglycemia and the development of ketoacidosis. Glucose homeostasis in these patients has been proposed to follow a circadian pattern (24), as it improves at night. In our patient, glucose levels remained abnormal even at night, when he was receiving continuous feeding, suggesting that fasting, rather than circadian variations, could have been responsible for the observed decreased glucose levels at night. Contrainsular hormones surely play a role in gluconeogenesis, and fluctuations in their levels have been associated with changes in glucose concentration (22). Hypophysectomy improves metabolic control in Rabson-Mendenhall syndrome, although only transiently (22). Direct measurements have indicated normal to elevated GH concentrations in these patients (24, 33, 34, 35), including ours (12). GH is, however, unable to increase circulating IGF-I levels (12, 24) and has only a minimal effect on fasting glucose levels in these patients (12). Measurement of cortisol and thyroid hormone have been normal in the patients studied (12, 21). Therefore, the changes in insulin levels described in this report are likely to be the major causative factor for the natural history of Rabson-Mendenhall syndrome.
The progressive decline in insulin levels mimics at a faster rate what is observed in patients with type 2 diabetes, in whom a decrease in insulin levels follows initial hyperinsulinemia (36, 37, 38). The changes observed in Rabson-Mendenhall syndrome in a few years probably require decades in type 2 diabetes and usually do not progress, except in rare cases, to an insulin-deficient state severe enough to cause ketoacidosis. Even in Rabson-Mendenhall syndrome there is variability in the rate of decline in insulin levels. Serrano Rios et al. observed residual excessive insulin release (measured as C peptide) in a 16-yr-old patient with Rabson-Mendenhall syndrome (39). Insulin levels were also much higher in this patient (1300–2200 µU/mL) than in our patient at the time of ketoacidosis, but this was probably due to the concomitant insulin therapy (39). This patient presented with diabetes for the first time at 11 yr of age and had a much milder phenotype than that of our patient. It is likely that the heterogeneity of mutations in the insulin receptor gene causing Rabson-Mendenhall syndrome explain the different clinical courses.
Our results also demonstrate that insulin resistance alone is not sufficient to cause full-blown diabetes in humans. Although patients with inherited insulin-resistant syndromes have abnormal glucose homeostasis from birth, they have both hypo- and hyperglycemia and not constant hyperglycemia as patients with type 2 diabetes. Only with time did our patient with Rabson-Mendenhall syndrome progress to constant hyperglycemia after insulin levels completed a phase of rapid decrease. The requirements for both insulin resistance and decreased insulin levels to achieve frank diabetes are also emerging from the study of knockout animal models (25, 40). In mice, knock out of the IRS-1 gene causes growth retardation and insulin resistance, which are, however, compensated by increased insulin secretion (41). A similar picture is observed in mice heterozygous for a knockout allele for the insulin receptor (10). However, overt diabetes develops with high frequency in IRS-2 knockout mice, whose absence impairs not only insulin action but also pancreatic ß-cell function (42), or mice double heterozygotes for impaired IRS-1 and glucokinase genes (43). Interestingly, tissue-specific knockout of the insulin receptor gene in pancreatic ß-cells creates an insulin secretory defect in mice (44). These animal data suggest that in our patient, a single genetic lesion may be responsible for both the extreme insulin resistance observed since birth and the progressive decline in insulin levels. The slower rate of decrease in insulin levels in the mouse model compared to our patient is probably due to the lack of peripheral insulin resistance caused by the presence of normal insulin receptors in extrapancreatic tissues (44).
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received March 24, 1999.
Revised April 28, 1999.
Accepted May 7, 1999.
|
References |
|---|
|
TopAbstractIntroductionExperimental SubjectsMaterials and MethodsResultsDiscussionReferences |
|---|
Tyr) in the insulin receptor
-subunit. Pediatr Res. 42:72–77.[Medline]
This article has been cited by other articles:
 |
|
 |
|
  E. Cochran, J. R. Young, N. Sebring, A. DePaoli, E. A. Oral, and P. Gorden Efficacy of Recombinant Methionyl Human Leptin Therapy for the Extreme Insulin Resistance of the Rabson-Mendenhall Syndrome J. Clin. Endocrinol. Metab., April 1, 2004; 89(4): 1548 - 1554. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. NANDI, Y. KITAMURA, C. R. KAHN, and D. ACCILI Mouse Models of Insulin Resistance Physiol Rev, April 1, 2004; 84(2): 623 - 647. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Ando, T. Yatagai, K. Rokkaku, S. Nagasaka, S.-e Ishikawa, and S. Ishibashi Obesity Is a Critical Risk Factor for Worsening of Glucose Tolerance in a Family With the Mutant Insulin Receptor Diabetes Care, August 1, 2002; 25(8): 1484 - 1485. [Full Text]
|
|
|
|
 |
|
 N. Longo, Y. Wang, S. A. Smith, S. D. Langley, L. A. DiMeglio, and D. Giannella-Neto Genotype-phenotype correlation in inherited severe insulin resistance Hum. Mol. Genet., June 1, 2002; 11(12): 1465 - 1475. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Nakae, Y. Kido, and D. Accili Distinct and Overlapping Functions of Insulin and IGF-I Receptors Endocr. Rev., December 1, 2001; 22(6): 818 - 835. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
