Diabetologia (1993) 36:1288-1292 Diabetologia Springer-Verlag 1993

The development of hyperglycaemia in patients with -resistant generalized lipoatrophic syndromes

J. J. Robert 1, B. Rakotoambinina 1, I. Cochet e, V. Foussier I , J. Magre e , D. Darmaun 3, D. Chevenne 4, J. Capeau 2 Unit of Pediatric Endocrinology and Diabetology, INSERM U 30, Necker-Enfants Malades Hospital, Paris, France 2Laboratory of Cellular Biology and Biochemistry,INSERM U 181, Saint Antoine Faculty of Medicine, Paris, France 3INSERM U 290, Saint-Lazare Hospital, Paris, France 4Laboratory of Physiology,Necker-Enfants Malades Hospital, Paris, France

Summary. is present in patients suffering duction, measured by infusion of [6,6-2H] glucose, varied from lipoatrophic syndromes long before the onset of from 1.7 to 8.3 mg/kg, rain and was significantly correlated mellitus. Thus, the decreased peripheral glucose dis- with fasting plasma glucose. The overproduction of glucose posal may not be the only mechanism of hyperglycaemia. despite basal hyperinsulinism suggested hepatic insulin re- The kinetic parameters of glucose homeostasis were evalu- sistance, which was confirmed by the abnormal response to ated in six young females aged 15,16, 18, 19 and 24 years with constant unlabelled glucose infusion (2 mg/kg, min) in five generalized lipoatrophy; one patient was studied both at 12 patients. In conclusion, impaired glucose tolerance seems to and 15 years. Insulin resistance was evaluated in vivo by the develop in generalized lipoatrophy with aggravated periphe- hyperinsulinaemic euglycaemic clamp (3M insulin infusion ral insulin resistance. The present data show that fasting rates from 1 to 100 mU/kg, min). All patients showed a right- hyperglycaemia is mainly the consequence of increased he- ward shift of the dose-response curve, indicating decreased patic glucose production. insulin sensitivity. In two patients, maximal glucose disposal was moderately decreased, while in five patients it was dra- Key words: Generalized lipoatrophic diabetes, insulin resis- matically reduced (3.6-6.9 mg/kg, min). Fasting plasma glu- tance, hyperinsulinaemic euglycaemic clamp, stable isotope cose was variable (4.3-18.3 mmol/1) and did not correlate labelled glucose, hepatic glucose production. with peripheral glucose disposal rates. Hepatic glucose pro-

Generalized lipoatrophy is a syndrome characterized by creased hepatic glucose release in those syndromes where the total disappearance of subcutaneous adipose tissue insulin resistance appears as the primary defect. However, [1]. Although rare, it is well defined as a clinical entity as the syndrome of generalized lipoatrophy is extremely among the heterogeneous syndromes of extreme insulin rare, this hypothesis has not been tested. resistance [2]. The generalized absence of fat is usually as- Many studies have clearly documented insulin resis- sociated with insulin-resistant diabetes mellitus and tance in vivo in this syndrome [3,8-16], without clearly hypertriglyceridaemia with consequent hepatomegaly [1]. identifying the organ(s) responsible. In a previous report Insulin resistance generally occurs early in life, mainly as using the euglycaemic clamp and stable isotope dilution isolated hyperinsulinism [3]. Glucose tolerance gradually methods along with a panel of in vitro investigations [17], deteriorates to reach, often during or after puberty [3], the we demonstrated that insulin resistance concerned not state of diabetes with hyperinsulinism and lack of ketosis. only peripheral glucose disposal, but also hepatic glucose As insulin resistance actually exists long before the release. However, there appeared to be no correlation be- onset of diabetes [3], it may not be the only mechanism tween the severity of the hyperglycaemic syndrome and underlying the development of sustained hyperglycaemia. either the degree of peripheral insulin resistance, the rate In another clinical syndrome, Type 2 (non-insulin-de- of hepatic glucose production, or the in vitro alterations at pendent) diabetes, where insulin resistance plays a major the receptor or the post-receptor levels. We have now ex- role, postabsorptive hyperglycaemia mainly depends on tended our study to three new patients and reinvestigated the overproduction of glucose by the liver [4-6], and it is a previously studied individual. These new data demon- not closely correlated with the degree of insulin resistance strate the predominant role of the liver in the fasting or insulin-stimulated glucose uptake [7]. Thus, hypergly- hyperglycaemia found in this extreme insulin-resistant caemia may also, at least partly, be the consequence of in- state. J. J. Robert et al.: Generalized lipoatrophic syndromes 1289 Table 1. Selected clinical characteristics of six patients with gener- In vivo methods' alized lipoatrophy W ZT DR HD1 JB ZO HD2 A 3-h oral (OGTF; 45 g/m 2) was performed a few days before the protocol described above. Insulin resistance was Fasting plasma glucose assessed by the euglycaemic damp technique, using multiple insulin (mmolf!) 4.3 5.0 5.0 8.9 9.4 15.5 18.3 infusion rates [18]. The clamp was performed using an automated glucose-controlled insulin infusion system (Glucose Controller Bio- HbAlc(%) a 4.7 5.6 6.4 7.5 10.9 12.5 12.4 stator; Life Science Instruments, Miles Lab. Inc., Elkhart, Ind., Age (years) 16 19 18 12 15 24 15 USA). The Biostator was connected to the patient in the afternoon preceding the clamp study, as previously described [17], and plasma Age at diagnosis (years) 2 16 18 7 < 1 23 - glucose was normalized. The insulin infusion rate needed to main- Height(cm) 164 146 155 137 167 162 144 tain plasma glucose at normal values during the overnight period helped determine the rates of continuous insulin infusion during the Weight (kg) 83 32 55 27 58 58 38 clamp. For the first three patients, it was 1,2 and 10 mU/kg min. For Cholesterol(mmol/1) b 3.3 4.1 4.0 3.8 5.2 6.1 4.3 the following studies, it was 5, 10, 20 or 100 mU/kg, min. The rate of glucose infusion was continuously adjusted by the Biostator so as to Triglyceride(mmol/1)c 2.3 1.0 4.9 0.5 2.0 9.2 1.6 maintain euglycaemia; values of the last 30 min of each 120-min pla- Metabolic values were those at the time of the study. More severe al- teau were used to draw the dose-response curve between plasma in- terations of lipid metabolism were observed in several patients dur- sulin levels and glucose infusion rates. Blood samples were taken at ing the clinical course of the disease. 10-min intervals throughout the clamp for determination of plasma Reference values for HbA1c: 4.9 _+ 0.7 %. free insulin. b Reference values for cholesterol: 4.7 + 1.6 mmol/1. Glucose turnover was determined, on another day, by infusion of c Reference values for triglycerides: 1.1 + 0.6 mmol/1 D-[6,6-ZH]glucose (98 % 2H, CEA, Gif-sur-Yvette, France) while plasma glucose was either normo- or hyperglycaemic, depending on the patient's status. Baseline blood samples were obtained to deter- Subjects and methods mine the basal, natural abundance level of isotopic enrichment in plasma glucose. A primed, continuous infusion of [6,6-2H]glucose was then given for 270 min at a rate of 0.04 mg/kg, min, using a cali- Patients brated syringe pump (Perfusor IV; B. Braun, Melsungen, Germany). Blood samples were collected between 150 and 180 min of infusion The patients were six unrelated adolescent females with gener- and used to calculate basal hepatic glucose production [19], while the alized lipoatrophy. Our previous publication [17] was concerned patient was at a spontaneous postabsorptive plasma glucose level. with three: ZT (19 years old), JB (15 years old) and HD1. HD1 was Unlabelted 10 % dextrose was infused at a rate of 2 mg/kg, rain from 12 years old and had moderate hyperglycaemia at the time of the 180 to 270 min [20] using a calibrated peristaltic pump. The deut- first study; she was examined again at age 15 years (HD2), at which erated glucose enrichment in the plasma between 210 and 270 min time she had severe hyperglycaemia (Table 1). DR, who was 18 was used to evaluate the endogenous glucose production in response years old and normoglycaemic, had presented i year earlier with to the infusion of exogenous glucose [20]. severe hypertriglyceridaemia complicated by acute pancreatitis; All blood samples were immediately centrifuged at 4~ and she suffered from marked hyperandrogenism with hirsutism, mus- plasma was separated and kept frozen at -20~ until analysed. cular hypertrophy and dysmenorrhea at the time of consultation. The concentrations of glucose in plasma and infusates were ZO was 24 years old and severely hyperglycaemic; she presented measured by the glucose-oxidase method (Autoanalyzer II; Beck- with marked hyperandrogenism; her brother has Type 1 (insulin- man, Fullerton, Calif., USA). Free and total immunoreactive dependent) diabetes, and she had been treated with insulin for insulin was measured using the method of separation with about 1 year. VV was 16 years old and normoglycaemic; she had polyethylene glycol and radioimmunoassay with separation on been followed since the age of 2 years for lipoatrophy, advanced charcoal (1NSIK-3 CEA kit; CIS Bio International, Gif-sur-Yvette, statural growth, advanced puberty and early and severe signs of France). Immunoreactive C-peptide was determined by a double hyperandrogenism (hirsutism and dramatic muscular hyper- antibody method (RIA-mat C-peptide kit; CIS). The isotopic en- trophy). Two patients were black Africans (JB, HD), two were In- richment of plasma glucose was determined using the butylboro- dians (DR was from Reunion Island and VV from Martinique), and nate-acetate derivatives [19] which were analysed by combined gas two were North African Caucasians (ZO, ZT). Four patients (JB, chromatography-mass spectrometry on a Nermag R1010T instru- HD, ZO, ZT) had a history of parental consanguinity. All the pa- ment (Nermag, Rueil, France). The parameters of glucose kinetics tients suffered from generalized lipoatrophy, muscular hyper- were calculated from steady-state isotope dilution principles [20]. trophy, and moderate hepatomegaly. Abnor- When plasma glucose is in steady state: malities of carbohydrate metabolism were diagnosed at various ages (Table 1). Two patients, HD and ZO, 'were treated with in- sulin, with doses of up to 360 and 390 IU/day, respectively. The in- sulin treatment was interrupted at least 24 h before the tests. The patients were not taking other medication at the time of the study. where R, is the turnover of glucose, iis the tracer infusion rate, Ei and Routine blood counts, serum electrolytes, creatinine and liver func- Ep are the labelled glucose enrichment in the infusate and the plas- tion tests were within normal limits. All patients were fed a diet ma, respectively. In the basal state, the turnover of glucose equals the containing sufficient amounts of complex carbohydrate before the hepatic glucose production. During infusion with unlabelled glucose study. at a constant rate, ira is equal to the sum of endogenous production Seven lean, healthy individuals (five men, two women) with and exogenous glucose infusion. Therefore, the hepatic glucose pro- no family , aged 25 + 2 years (22-29 years), were duction rate is calculated by subtracting the known rate of infusion of used as a control group. Three different 2-h plateaus were ob- unlabelled glucose from the calculated R,. tained in each subject, using insulin infusion rates from 0.2- 5 mU/kg, min. The patients and their parents and individuals in the control group were fully informed of the purpose of the investi- gation, and they gave their consent to participate. The study was Statistical analysis approved by the ethical committee of Necker-Enfants Malades Hospital. Data were analysed with correlation coefficients when appropriate. 1290 J. J. Robert et al.: Generalized lipoatrophic syndromes

Results Conversely, there was a correlation (r = 0.90, p < 0.01) between fasting plasma glucose and basal hepatic glucose In vivo studies production (Fig.2). Hepatic glucose production was in- creased despite high fasting plasma insulin and C-peptide The results of the OGTTs for the first three patients have levels (Table 2). There was no significant correlation be- been previously reported [17]. They were clearly abnor- tween fasting plasma glucose or hepatic glucose produc- mal, but varied greatly between patients. OGTTs were not tion and either plasma insulin, C-peptide or non-esterified performed on two of the new cases, ZO and HD2, as they fatty acids, cholesterol and triglycerides (Tables 1 and 2). were severely hyperglycaemic. For patient DR, who was Hepatic glucose production was 2 mg/kg.min lower normoglycaemic in the fasting state, plasma glucose than in the basal state in only one patient, during the infu- reached a maximum of 12.4 mmol/1 at 90 min of the test, sion of glucose at 2 mg/kg, min. In the other patients, in- and returned to 7.2 mmol/1 at 180 min; plasma insulin was 2438 pmol/l at 90 min. Patient VV was normoglycaemic in 20 the fasting state and peaked at 9.7 mmol/1 at 30 min of the _'g OGTT, returning to 5.8 mmol/1 at 120 min; plasma insulin E reached 5449 pmol/1 at 120 min. HbAlo (normal values: ~ 15 4.9 + 0.7 %, mean + SD) was correlated with fasting plas- ...... 1" ma glucose (Table 1, p < 0.01). | The amount of insulin needed to keep the patients ~ to. normoglycaemic overnight, before the clamp was started, g varied from 0 to 5.3 mU/kg, min; the highest values (4.2 '4 7/" and 5.3 mU/kg, min) were observed for the two severely .--- hyperglycaemic patients. All patients showed a rightward shift of the dose-response curve to insulin for glucose in- _o= fusion rate, compared to values established in seven ca o young adult individuals (Fig. l), indicating that less 0 50 t00 1000 10000 glucose was disposed of by the tissues at comparable Plasma insulin (mU/I) plasma insulin levels. In two patients, ZT and HD1, Irig.l. Dose-response curves of insulin action on glucose disposal glucose disposal reached 9.0 and 9.2 mg/kg, min respec- rates. Correlations between glucose infusion rates and plasma in- tively with an insulin infusion rate of 10 mU/kg.min sulin levels during hyperinsulinaemic euglycaemic clamps per- (plasma insulin at 3485 and 2330 pmol/1, respectively), formed on six patients with generalized lipoatrophy ( VV; -k ZT; while for the healthy subjects the glucose disposal rate DR; HDI; JB; * ZO; O HD2), and for seven young adult was always greater than 12.2 mg/kg-min with insulin in- healthy subjects ( fusions of 5 or 10 mU/kg- min. Higher rates were not in- fused in these, the first two patients. The maximal glu- cose disposal rate was dramatically reduced to between 3 20 .... Y and 4 mg/kg, rain in four cases, and to 6.9 mg/kg- min in ....,..Y" | Fig.2. Correlation be- one patient, despite insulin infusion rates of up to E 12 ...... " tween fasting plasma ..y" glucose concentration 100 mU/kg, min. Two of these five patients, ZO and .- ..Z" HD2, were those with the most severe hyperglycaemia, ~ 10 and hepatic glucose e- production in six pa- but one was a patient with moderate fasting hypergly- '~ .Y 5 . x...~ . tients with gener- caemia, JB, and two, DR and VV, were normoglycaemic. alized lipoatrophy Thus, there was no correlation between the degree of in- ~. (A VV; * ZT; sulin resistance for glucose disposal and the postabsorp- o ; lb DR; HD1; JB; tive plasma glucose levels. Hepatic glucose production (mg/kg min) ZO; O HD2)

Table 2. Hepatic glucose production (HGR mg/kg -rain) and insulin metabolism in six lipoatrophic diabetic patients in the postabsorptive state and during exogenous glucose infusion (2 mg/kg, min) W ZT DR HD1 JB ZO HD2 Basal Plasma glucose (mmol/1) 4.3 5.0 5.0 8.9 9.4 15.5 18.3 HGPa 1.7 3.3 2.2 3.4 2.6 4.9 8.3 Plasma insulin (pmol/1)b 401 283 513 373 270 158 767 Plasma C-peptide (pmol/1)c 1900 1010 2250 564 935 735 580 Plasma NEFA (mmol/l) d 0.38 1.42 1.40 2.44 - 0.88 1.66 Glucose infusion HGP suppression - 1.5 1.7 1.5 2.0 0.6 1.6 a Values for HGP in control subjects varied from 1.7 to c Control values for plasma C-peptide: 570 _+ 130 pmol/1. 3.3 mg/kg, min, with a mean value of 2.4 + 0.1 mg/kg-min. d Control values for plasma NEFA: 0.29 _+0.27 mmol/1 b Control values for plasma insulin: 54 + 39 pmol/1. J. J. Robert et al.: Generalized lipoatrophic syndromes 1291 cluding the two with normoglycaemia, it was less sup- fasting plasma glucose level does not depend primarily on pressed (Table 2). The extent to which hepatic glucose the severity of peripheral insulin resistance in these syn- production was suppressed did not correlate with fasting dromes of extreme insulin resistance. Insulin resistance is plasma glucose, basal hepatic glucose production or plas- principally responsible for the postprandial glucose in- ma insulin levels. Plasma insulin and C-peptide remained tolerance. Diabetes develops permanently as a conse- close to the basal level during glucose infusion; insulin in- quence of increased production of glucose by the liver, creased from 384 +79 (mean + SEM) to 411 + 78 pmol/1 which occurs despite high circulating insulin levels. and C-peptide from 1012 + 236 to 1147 + 253 pmol/1. The predominant role of hepatic glucose production in the development of hyperglycaemia in the other types of diabetes has been well established. The overproduction of Discussion glucose is completely responsible for the fasting hypergly- caemia in [21]. It has also been shown that Our previous report on three patients with generalized hepatic insulin resistance, as manifested by alterations in lipoatrophy [17], indicated insulin resistance at several glucose production, may be the predominant factor con- organ sites: muscle, liver and islet beta cells. The inclusion tributing to fasting hyperglycaemia in [5, of three more patients and the reinvestigation of a pre- 6]. In the syndromes of extreme insulin resistance, such as viously studied individual has given a much clearer picture generalized lipoatrophy, fasting hyperglycaemia seems to of the mechanisms of hyperglycaemia in these syndromes develop following the same alterations in the kinetic pa- of insulin resistance. rameters of glucose metabolism, although the primary The hyperinsulinaemic euglycaemic clamp showed mechanisms are different. marked insulin resistance in all patients, with a rightward However, the present observations clearly show that shift of the dose-response curve between plasma insulin the relationship between insulin resistance at the hepatic levels and glucose disposal rates. However, two of the six and peripheral levels is not as simple as that seen in Type 2 patients had a slightly decreased capacity to utilize glu- diabetes [6]. Patients with the same extreme peripheral in- cose; they were normo- and moderately hyperglycaemic sulin resistance showed varying capacities to control he- in the postabsorptive state. Two other patients, who were patic glucose output. In addition, the levels of circulating comparable to the first two in terms of fasting blood glu- insulin required to maintain the fasting plasma glucose at cose, glucose intolerance and HbAlc, had a very low maxi- normal levels varied greatly, as did the response of hepatic mal capacity for glucose disposal, even at very high plasma glucose production to increased endogenous insulin se- insulin concentrations. Finally, two patients were severely cretion caused by the infusion of a low dose of glucose. hyperglycaemic and extremely insulin resistant. One pa- None of these various parameters were correlated. In ad- tient, who was studied twice with a 3-year interval, de- dition, differences in the lipid status of the patients did not veloped hyperglycaemia while insutin resistance had seem to be related to the peripheral insulin resistance or greatly increased. Although these observations suggest hepatic glucose production. that hyperglycaemia and peripheral insulin resistance In vitro studies, performed in the six patients, are in may develop concomitantly, fasting plasma glucose varied good agreement with results published previously [17, 22, greatly for the same severity of insulin resistance. 23]. The four patients who had never been treated with in- Conversely, plasma glucose was strongly correlated sulin had no insulin autoantibodies, and none had insulin with the rate of hepatic glucose production. As reported receptor autoantibodies. Insulin binding, studied on ery- previously [17], there was no detectable difference be- throcytes and on skin fibroblasts in primary cultures, was tween normo- and moderately hyperglycaemic patients. either normal (VV, HD1, JB, ZT and ZO) or about 50 % This is not surprising, since variations of hepatic glucose sub-normal (DR). At the post-binding level, there were production in the 5-10 mmol/1 range have been shown to alterations in the tyrosine kinase activity for all patients, be small in other types of diabetes. They are less than the concerning either the autophosphorylation process or the inter-individual variations and can be identified only by exogenous tyrosine kinase activity or both [23]. Similar comparing the same individual at different plasma glucose conclusions were drawn concerning the post-receptor in- levels [21]. Correlations with hepatic glucose production sulin action on glucose metabolism (hexose transport or become readily measurable only when the plasma glucose glycogen synthesis), while the activation of amino acid levels are higher than approximately 10 mmol/1 [4-6, 21]. transport by insulin was unaffected [22]. Thus, defects Although these observations were made on a small were present at the cellular level, resulting in insulin re- number of subjects, they are in agreement with other sistance in vitro in all cases. The genetic defects respon- studies of rare individuals with generalized lipoatrophy. sible for this resistance are still unknown [24, 25]. The pat- Increased hepatic glucose production has been correlated tern and extent of the alterations were heterogeneous with fasting hyperglycaemia in two patients [13, 14], and in both in vivo and in vitro. However, there were no clear one it returned to normal along with the normalization of correlations between the in vivo and the in vitro data. This glycaemia following dietary and insulin therapy [13]. In a is not surprising, since in vivo results partly reflect alter- recent study of another type of extreme insulin resistance ations secondary to the metabolic and hormonal status with acanthosis nigricans, four patients with diabetes had [26, 27]. This was clearly shown for the patient who was much greater hepatic glucose production than did the studied on two occasions. Insulin resistance was moderate non-diabetic patients, while the peripheral insulin resis- and insulin binding to erythrocytes was normal during the tance was only slightly greater [16]. This suggests that the first study. During the second investigation, when fasting 1292 J. J. Robert et al.: Generalized lipoatrophic syndromes

blood glucose had dramatically worsened, both param- 13. Keenan BS, Kirkland RT, Garber AJ et al. (1980) The effect of eters were markedly impaired. diet upon carbohydrate metabolism, insulin resistance, and blood pressure in congenital total lipoatrophic diabetes. Metabo- In conclusion, although we have shown that the bio- lism 29:1214-1224 logical expression of insulin resistance at its various sites 14. Beylot M, Sautot G, Laville M, Cohen R (1988) Metabolic appears to be variable, the nature of their interrelation- studies in lipoatrophic diabetes: mechanisms of ships should be further explored to provide a clearer pic- and evidence of resistance to insulinof lipid metabolism. Diabete ture of the progressive development of hyperglycaemia in Mdtab 14:20-24 syndromes of extreme insulin resistance, such as gener- 15. Tsukahara H, Kikuchi K, Kuzuya H et al. (1988) Insulin resis- tance in a boy with congenital generalized . Pediatr alized lipoatrophy. Res 24:668-672 16. Cohen R Harel C, Bergman R et ai. (1990) Insulinresistance and Acknowledgements. We are indebted to Professors C. De Gennes acanthosis nigricans: evidence for a postbinding defect in vivo. (H6pital de la Piti& Paris), J. C.Dayras (H6pital Delafontaine, St- Metabolism 39:1006-1011 Denis), E Kutten (H6pital Necker-Enfants Malades, Paris), and 17. Robert JJ, Magre J, Reynet C, Darmaun D, Picard J, Capeau J Drs. B. Le Luyer (Centre Hospitalier, Le Havre), R Tric (H6pital de (1990) In vivo and in vitro characterization of insulin resistance la Cit6 Universitaire, Paris), J.Tlapin (Centre Hospitalier, Meaux), in three cases of fipoatrophic diabetes. Diab~te Mdtab 16: for giving us the opportunity to study these patients. We thank Dr. 240-247 E Grigorescu for performing erythrocyte insulinbinding studies. We 18. Rizza RA, Mandarino LJ, Gerich JE (1981) Dose-response char- are grateful to Ms M. J. Blivet for excellent technical assistance. We acteristics for effects of insulin on production and utilization of are extremely grateful to the nursing staff of the Metabolic Unit for glucose in man. Am J Physio1240:E630 E639 their support in carrying out these demanding studies. This work was 19. Bier DM, Arnold KJ, Sherman WR, Holland WH, Holmes WE supported by grants from the Institut National de la Sant6 et de la Kipnis DM (1977) In vivo measurement of glucose and alanine Recherche M6dicale, the Aide aux Jeunes Diab6tiques, and the Con- metabolism with stable isotopic tracer. Diabetes 26:1005-1015 seil Scientifique-Facult6 de M6decine Saint Antoine, Paris, France. 20. Wolfe RR, Allsop JR, Burke JF (1979) Glucose metabolism in man: responses to intravenous glucose infusion. Metabolism 28: 210-219 References 21. Darmaun D, Cirillo D, Koziet J, Chauvet D, Young VR, Robert JJ (1988) Whole body glucose kinetics in type I diabetes studied 1. Lawrence RD (1946) Lipodystrophy and hepatomegaly with with (6,6-2H) and (U-13C) glucose and the artificial B-cell. Me- diabetes, lipaemia, and other metabolic disturbances: a case tabolism 37:491-498 throwing light on the action of insulin. Lancet I: 724-731 22. Magre J, Reynet C, Capeau J, Blivet MJ, Picard J (1988) In vitro 2. Kahn CR, Flier JS, Bar RS et al. (1976) The syndromes of insulin studies of insulin resistance in six patients with lipoatrophic resistance and acanthosis nigricans. N Engl J Med 294:739-745 diabetes: evidence for heterogeneous post-binding defects. 3. Oseid S (1973) Studies in congenital generalized lipodystrophy Diabetes 37:421-428 (Seip-Berardinelli syndrome). Acta Endocrino172:475-494 23. Magre J, Grigorescu E Reynet C et al. (1989) Tyrosine-kinasede- 4. B owen HF, Moorhouse JA (1973) Glucose turnover and disposal fect of the insulin receptor in cultured fibroblasts from patients in maturity-onset diabetes. J Clin Invest 52:3033-3045 with lipoatrophic diabetes. J Clin Endocrinol Metab 69:142-150 5. De Fronzo RA (1988) The triumvirate: fl-cell, muscle, liver. A 24. Van der Vorm ER, Kuipers A, Bonenkamp JW et al. (1993) Pa- collusion responsible for IDDM. Diabetes 37:667~587 tients with lipodystrophic diabetes mellitus of the Seip-Berardi- 6. Campbell P J, Mandarino L J, Gerich JE (1988) Quantification of nelli type, express normal insulin receptors. Diabetologia 36: the relative impairment in actions of insulin on hepatic glucose 172-174 production and peripheral glucose uptake in non-insulin-de- 25. Desbois C, Magr6 J, Amselem S, Goossens M, Capeau J, pendent diabetes mellitus. Metabolism 37:15-21 Besmond C (1993) Lipoatrophic diabetes: exclusion of the in- 7. Reaven GM (1988) Role of insulin resistance in human disease. sulin-receptor gene. Exp Clin Endocrino1101 [Suppl]: 352-354 Diabetes 37:1595-1607 26. Klip A, Pfiquet MR (1990) Glucose transport and glucose trans- 8. Ruvalcaba RH, Samols E, Kelley VC (1965) Lipoatrophic porters in muscle and their metabolic regulation. Diabetes Care diabetes. I. Studies concerning endocrine function and carbohy- 13:228-243 drate metabolism. Am J Dis Child 109:279-286 27. Farese RV, Standaert ML, Arnold T et al. (1992) The role of pro- 9. Samaan NA, Craig JW (1969) Serum insulin and growth hor- tein kinase C in insulin action. Cellular Signalling 4:133-143 mone in lipoatrophic diabetes. Metabolism 18:460-468 10. Senior B, Loridan L (1969) Fat cell function and insulin in a pa- Received: 16 March 1993 tient with generalized lipoatrophy. J Pediatr 47:972-975 and in revised form: 29 June 1993 11. Naito C, Togawa K (1974) A possible role of circulating lipopro- tein-triglycerides in the increase in concentration of free fatty Professor J.-J. Robert acids and in insulinresistance in "total" lipodystrophy. J Clin En- Unit6 d'Endocrinologieet Diabdtologie de 1'Enfant docrinol Metab 39:1030-1037 HSpital des Enfants Malades 12. Huseman C, Johanson A, Varma M, Blizzard RM (1978) Con- 149 rue de S6vres genital lipodystrophy: an endocrine study in three siblings. I. Dis- F-75015 Paris orders of carbohydrate metabolism. J Pediatr 93:221-226 France