Type 1 Diabetes Associated with Acquired Generalized Lipodystrophy and Insulin Resistance: The Effect of Long-Term Leptin Therapy

Abstract

Context: Acquired generalized lipodystrophy (AGL) is marked by severe insulin resistance and hypertriglyceridemia. Rarely, AGL and type 1 diabetes (T1D) coexist.

Objective: Our objective was to describe the response to leptin therapy in patients with coexisting AGL and T1D and to document the autoimmune diseases associated with AGL.

Design and Setting: We conducted an open-label prospective study at the Clinical Research Center of the National Institutes of Health.

Patients: Participants included 50 patients with generalized or partial lipodystrophy (acquired or congenital); two patients had both AGL and T1D.

Intervention: Patients were treated with 12 months of recombinant human leptin administration to achieve high-normal serum concentrations.

Results: Two patients had both AGL and T1D. The first was diagnosed with T1D at age 8 yr. Beginning at age 11 yr, he developed generalized lipodystrophy, elevated transaminases, and poor glycemic control [hemoglobin A_1c (HbA_1c) 10.7%] despite markedly increased insulin requirements (3.3–5 U/kg·d). Further evaluation revealed hypoleptinemia and hypertriglyceridemia. At age 15 yr, leptin therapy was initiated, and after 1 yr, his insulin requirements fell to 1 U/kg·d, his glycemic control improved (HbA_1c 8.4%), and both his triglycerides and transaminases normalized. The second patient developed concurrent AGL and T1D at age 6 yr. Despite insulin doses of up to 32 U/kg·d, she developed poor glycemic control (HbA_1c 10.6%), hypertriglyceridemia (2984 mg/dl), elevated transaminases, and nonalcoholic steatohepatitis. At age 13 yr, leptin therapy was started, and after 1 yr, her glycemic control improved (HbA_1c 7.3%) and her insulin requirements decreased (17 U/kg·d). Her triglycerides remained elevated but were improved (441 mg/dl).

Conclusions: Long-term recombinant leptin therapy is effective in treating the insulin resistance of patients with the unusual combination of T1D and AGL.

Leptin therapy on two patients with acquired generalized lipodystrophy and type 1 diabetes was found to improve their glycemic control, decrease their insulin requirements, and improve their triglycerides, which suggests that this treatment is effective for these two disorders.

Generalized lipodystrophy represents a clinical condition that is characterized by widespread loss of adipocyte tissue, extreme insulin resistance, diabetes, dyslipidemia, and nonalcoholic steatohepatitis. The condition may have a genetic or acquired etiology, but the major therapeutic challenge in either case is treatment of the metabolic dysfunction and insulin resistance (1). We and others have demonstrated that therapy with the adipocyte-derived hormone leptin ameliorates the metabolic derangements of generalized lipodystrophy (2,3,4,5).

Type 1 diabetes (T1D) is an autoimmune disease marked by T cell-mediated destruction of pancreatic β-cells culminating in absolute insulin deficiency. When T1D occurs in a patient with generalized lipodystrophy, a new phenotype emerges that is characterized by extreme insulin resistance and severe insulinopenia, rather than the hyperinsulinemia typically seen with insulin resistance. Thus, the hallmark of the new phenotype is glycemic lability superimposed on insulin resistance and associated dyslipidemia.

The etiology of acquired partial and generalized lipodystrophy is unknown, but it may be autoimmune in nature (6,7). Many of the patients have a history of panniculitis that precedes the development of lipodystrophy (6,8). Furthermore, acquired lipodystrophy may be associated with autoimmune diseases (6,7,9).

In the present study, we describe two patients with acquired generalized lipodystrophy (AGL) who have T1D. We further demonstrate the role of leptin in the management of this condition. In addition, we show that autoimmune conditions are common in our patients with acquired lipodystrophy.

Patients and Methods

Patient population

We evaluated a total of 50 patients with either generalized or partial lipodystrophy (acquired or congenital). Two of these patients had both AGL and T1D.

Recombinant methionyl human leptin (r-metHuLeptin) therapy was given as a self-administered, twice-daily sc injection as previously described (5). The dose was escalated to the full dose over the first 2 months of treatment. Thereafter, the usual replacement dose was 0.08 mg/kg·d in an attempt to simulate the normal to high physiological range. Patients were evaluated at the Clinical Research Center of the National Institutes of Health at baseline, every 4 months for 1 yr, and every 6 months thereafter. Inpatient data were collected on a metabolic unit during each visit. At baseline, patients were on conventional treatment for diabetes and dyslipidemia.

The protocol was approved by the institutional review board of the National Institute of Diabetes and Digestive and Kidney Diseases. Informed consent was obtained from each patient.

Biochemical analyses

Levels of serum leptin, hemoglobin A_1c (HbA_1c), glucose, and lipids were measured as previously described (3,4,5,10). C-peptide levels were measured with a two-site chemiluminescent enzyme immunometric assay on DPC Immulite 2000 equipment (Diagnostic Products Corp., Los Angeles, CA). Glutamic acid decarboxylase antibody (anti-GAD) was measured with a RIA (Mayo Medical Laboratories, Rochester, MN). ELISAs were used to detect the presence of thyroglobulin antibody and thyroid peroxidase antibody (Trinity Biotech, Jamestown, NY). Anti-nuclear antibody levels were measured with a qualitative enzyme immunoassay (Bio-Rad, Laboratories, Hercules, CA).

Procedures

Changes in calorie and macronutrient intake, resting energy expenditure, percent body fat, and liver volumes were assessed as previously described (3,4,5,10).

Results

Baseline characteristics

NIH-28

NIH-28, a Caucasian male, was diagnosed with T1D at 8 yr of age after he presented with mild ketoacidosis. At 10 yr of age, he developed mild panniculitis posterior to the right ear. At age 11 yr, his hyperglycemia worsened and his insulin requirements increased to 200–300 U/d (3.3–5 U/kg·d). At the same time, he was noticed to have developed more prominent musculature, a generalized loss of sc fat, and a voracious appetite. Further investigation revealed that he was hypoleptinemic and had elevations in his serum aminotransferases and triglycerides. Despite initiation of continuous sc insulin infusion therapy with basal rates in excess of 8 U/h, he was unable to achieve adequate glycemic control.

The patient was initially evaluated at the National Institutes of Health at age 14 yr and was found to have the following metabolic abnormalities at baseline: an extremely low serum leptin of 1.09 ng/ml (reference range 3.8 ± 1.8 ng/ml), an elevated HbA_1c of 10.7% despite being on high doses of insulin, high serum triglycerides of 282 mg/dl (Table 1), and elevated transaminases [alanine aminotransferase (ALT) 139 U/liter and aspartate aminotransferase (AST) 76 U/liter]. A liver biopsy showed glycogenosis, mild focal steatosis, and mild chronic hepatitis, but the pattern of injury was felt not to be that of nonalcoholic steatohepatitis (NASH). An anti-GAD level was elevated at 0.64 nmol/liter (reference range 0–0.02 nmol/liter), and a C-peptide level was undetectable at less than 0.5 ng/ml (reference range 0.9–4.0 ng/ml). The patient had persistently undetectable C-peptide levels on both fasting and oral glucose tolerance testing.

Table 1.

Changes in physical and metabolic parameters during r-metHuLeptin therapy

Test	Patient no.a	Baseline	4 months	8 months	12 months
Leptin (ng/ml)b	NIH-28	1.09	18.7	11	10.6
	NIH-9	3.95	15.19	4.04	11.62
HbA1c (%)	NIH-28	10.7	9	8.5	8.4
	NIH-9	10.6	7.9	9.6	7.3
Fasting glucose (mg/dl)	NIH-28	179	242	66	267
	NIH-9	298	603	558	253
C-peptide (ng/ml)	NIH-28	<0.5	<0.5	<0.5	<0.5
	NIH-9	1.4	1.7	0.9	0.6
Triglycerides (mg/dl)	NIH-28	282	96	53	86
	NIH-9	2984	4143	3695	441
Total cholesterol (mg/dl)	NIH-28	95	85	98	102
	NIH-9	374	686	406	115
HDL (mg/dl)	NIH-28	21	25	28	23
	NIH-9	20	20	18	11
LDL (mg/dl)	NIH-28	18	41	59	62
	NIH-9	165	270	210	38
Liver volume (ml)	NIH-28	2291.47	1597.1	1507	1719
	NIH-9	4291	4410	4628	3174
Weight (kg)	NIH-28	59.1	49.5	49.6	52.6
	NIH-9	31	31.4	31.8	32.8
BMI (kg/m2)	NIH-28	21.7	18	18	19.1
	NIH-9	15.4	15.5	15.6	15.6
Body fat (%)c	NIH-28	6.3	5.2	No study	5.9
	NIH-9	8.0	7.7	9.4	5.7
Energy expenditure (kcal/24 h)	NIH-28	2040	1180	1310	1290
	NIH-9	1690	1930	1620	1880

BMI, Body mass index; HDL, high-density lipoprotein; LDL, low-density lipoprotein. 

^aNIH numbers correspond to previous publications. 

^bNormal fasting range of leptin is 3.8 ± 1.8 ng/ml for males and 7.4 ± 3.7 ng/ml for females. 

^cPercent body fat values represent subtotal body fat measurements, which include the entire body contribution except the head. 

NIH-9

NIH-9, a Caucasian female, developed both AGL and T1D at 6 yr of age. Portions of her clinical course have been described previously (11). Her insulin requirements increased rapidly, and she continued to have poor glycemic control despite insulin doses of up to 1000 U/d (32 U/kg·d). The difficulty in achieving euglycemia reflected a combination of insulin resistance and tremendous glycemic lability as her hyperglycemia was punctuated by frequent occurrences of hypoglycemia. At the time of study enrollment, she was 13 yr old. She had a low C-peptide level of 1.4 ng/ml and was later found to have an elevated anti-GAD level of 123 nmol/liter. On repeated fasting and oral glucose tolerance testing, her C-peptide levels were either low or undetectable. Before the initiation of r-metHuLeptin therapy, her serum triglyceride level was 2984 mg/dl while on fenofibrate and pioglitazone. The patient also had elevated transaminases (ALT 79 U/liter and AST 85 U/liter), and a liver biopsy revealed evidence of NASH. More inflammation than expected with NASH was seen; the inflammation was more similar to that seen in autoimmune hepatitis and was associated with mild bridging fibrosis. The patient’s baseline state was also notable for a high urinary protein excretion of 2.8 g/24 h, a normal serum creatinine, and an elevated creatinine clearance of 174 ml/min (11).

Effects of recombinant leptin therapy

NIH-28

After 4 months of r-metHuLeptin therapy, the serum triglyceride levels of the patient normalized and remained in the normal range over his 12 months of follow-up (Fig. 1 and Table 1). Over the course of the first year of leptin therapy, his insulin requirements decreased to 50 U/d (1 U/kg·d) and his HbA_1c fell to 8.4% despite relatively poor compliance with basal-bolus insulin therapy. His liver function tests normalized with ALT and AST levels of 19 U/liter. A liver biopsy at his 12-month visit again showed mild chronic inflammation but complete resolution of his glycogenosis and steatosis. His pattern of improved metabolic control continued for a second year of leptin therapy. However, at the age of 17 yr, the patient was no longer under intense parental supervision and became increasingly noncompliant with both his leptin and insulin therapy. At that time, leptin was discontinued. The patient’s most recent HbA_1c and serum triglycerides were 13.4% and 3346 mg/dl, respectively.

Figure 1.

Clinical course of NIH-28 and NIH-9 on r-metHuLeptin therapy. Arrow with inset represents serum triglycerides of NIH-9 while on a 1-wk controlled high monounsaturated fat diet.

NIH-9

The clinical course for NIH-9 was somewhat more complex. While on r-metHuLeptin therapy, her insulin requirements and HbA_1c decreased dramatically by 4 months to 55 U/d (1.7 U/kg·d) and 7.3%, respectively (Fig. 1). Her hypertriglyceridemia, however, did not improve. Her glycemic control worsened 8 months into treatment, and her insulin requirements increased to 330 U/d (10.4 U/kg·d). She was then placed on a 1-wk controlled high monounsaturated fat diet, which had no effect on her triglyceride levels (Fig. 1). At that time, it was reported that lipoatrophic A-ZIP/F-1 mice treated with a peroxisome proliferator-activated receptor-γ (PPAR-γ) agonist actually had increased liver weight and hepatosteatosis (12). In light of these studies, pioglitazone was discontinued 10 months after initiation of leptin treatment when the patient had serum triglycerides of 3670 mg/dl (Fig. 1). Her insulin requirements rose with the cessation of pioglitazone to 550 U/d (17 U/kg·d) at the 12-month visit but were still below baseline values (1000 U/d). By the patient’s 12-month visit on leptin therapy, her hypertriglyceridemia had significantly improved with serum triglycerides of 441 mg/dl. In addition, her liver volume had decreased and her HbA_1c had fallen to 7.3% (Fig. 1 and Table 1).

On her liver biopsies at her 4-month and 12-month visits, the patient had progressively more inflammation that was not typical of NASH. Her liver laboratory values worsened over time such that her ALT and AST levels were as high as 445 U/liter and 2229 U/liter, respectively. They did not improve or stabilize with the discontinuation of pioglitazone at 10 months into leptin replacement or of leptin at 14 months.

Unfortunately, the patient’s preexisting proteinuria worsened after 14 months of leptin therapy, such that she was excreting over 15 g of protein a day. A renal biopsy later revealed membranoproliferative glomerulonephritis (MPGN) type 1 (11). This pathology differs from MPGN type 2, which is associated with hypocomplementemia and is seen in partial lipodystrophy (7,13). Leptin therapy was discontinued, but her massive proteinuria persisted for the following 8 months, and she ultimately died at age 15 of progressive renal failure (11).

In both patients, plasma leptin levels increased over the 12 months of r-metHuLeptin therapy (Table 1). Leptin treatment was associated with a decrease in weight, body mass index, resting energy expenditure, and liver volume in NIH-28 but had a somewhat more variable effect on the same parameters in NIH-9. Similar to our previous observations, there was no significant change in percent body fat or high-density lipoprotein in either patient (3).

Autoimmune diseases associated with lipodystrophy

Because autoimmune disorders have previously been noted in patients with acquired lipodystrophy (6,7,9), we sought to systematically determine the occurrence of coexisting autoimmune diseases in our patients (Table 2). We have noted for the first time the coexistence of T1D and immune markers of T1D in AGL. We further observe the frequent occurrence of autoimmune thyroiditis, dermatomyositis, and autoimmune hepatitis in our series of acquired lipodystrophy patients. By contrast, these conditions were rarely seen in our patients with congenital forms of lipodystrophy.

Table 2.

Autoimmune diseases in lipodystrophy

Type of lipodystrophy	Total no. of patients	T1D	Positive anti-GAD	Juvenile dermatomyositis	Autoimmune heptatitis	Celiac disease	Hashimoto’s thyroiditis
Acquired generalized	12	2	5	2	4		10
Congenital generalized	23					1	2
Acquired partial	6			4
Congenital partial	9

Discussion

In this report, we describe two patients with concomitant T1D and AGL. This combination creates a unique metabolic phenotype characterized by insulin resistance and glycemic lability and presents an unusual therapeutic challenge. The extreme insulin resistance associated with generalized lipodystrophy is usually marked by endogenous hyperinsulinemia and/or requirements of large doses of exogenous insulin. By contrast, T1D is characterized by near absent or markedly impoverished endogenous insulin secretion. We and others have shown that recombinant leptin is a novel therapeutic agent in the treatment of generalized lipodystrophy (2,3,5). The major effect of leptin therapy is to increase insulin sensitivity, reduce hypertriglyceridemia, remove ectopic fat, and modestly improve insulin secretion. In the present study of concurrent AGL and T1D, we show that leptin reduces insulin resistance and ectopic fat but has no effect on insulin secretion.

Although the two patients in our report have several similarities, important differences in their presentation and clinical course merit discussion. Both had hyperglycemia, high insulin requirements, and hypertriglyceridemia characteristic of insulin resistance. Both also demonstrated the glycemic lability associated with severe insulin deficiency. However, the degree of hypertriglyceridemia was nearly 10-fold higher in NIH-9. This disparity in the degree of hypertriglyceridemia may be explained by the phenotypic heterogeneity of lipodystrophy. Another difference between the two patients is that NIH-9 had more severe liver disease. Her liver dysfunction appeared to be caused by two distinct processes: her liver biopsies not only had features of NASH but also demonstrated progressive inflammatory change that was atypical of NASH and more characteristic of chronic autoimmune hepatitis.

Although NIH-9 represents only a single case, her drop in serum triglycerides after the discontinuation of pioglitazone is notable. When thiazolidinediones (TZDs) are used to treat patients with at least some fat tissue (e.g. patients with partial lipodystrophy or type 2 diabetes), an increase in sc fat is seen, possibly from increased adipocyte differentiation (14,15,16,17,18,19,20,21). The effect of TZDs and PPAR-γ activation in generalized lipodystrophy patients with complete absence of adipose tissue may be different. In these cases, PPAR-γ activation occurs mostly in the liver, resulting in unknown metabolic complications. In A-ZIP/F-1 mice, which have no white fat tissue and are models of generalized lipodystrophy, rosiglitazone worsens hepatosteatosis and increases liver weight (12). Although it is tempting to attribute the persistence of hypertriglyceridemia in NIH-9 to pioglitazone use, generalizations cannot be made at this point. However, on the basis of the rodent and human experience, we do not recommend the use of TZDs in patients with generalized lipodystrophy.

Autoimmunity may be a feature common to acquired lipodystrophy and T1D. Although an autoimmune basis for T1D is well accepted, the pathogenesis of acquired generalized and partial lipodystrophies is unclear. It has been hypothesized that acquired lipodystrophy represents an autoimmune disorder because it has been associated with other autoimmune diseases (6,7). Moreover, it has been proposed that the presence of other autoimmune diseases be considered a supportive diagnostic criterion for acquired lipodystrophy (6,7). In our series of patients, the presence of other autoimmune disorders was more common in those with acquired than congenital lipodystrophy. In a large review of AGL, associated autoimmune conditions were common, but there was no mention of the concurrence of T1D or immune markers of T1D (6). To our knowledge, our paper is the first report of coexisting AGL and T1D.

The relationship between the hypoleptinemia of lipodystrophy and immune dysfunction remains to be determined. Leptin has been implicated in the activation of autoimmunity. For example, leptin-deficient ob/ob mice that are immunized with a myelin-derived peptide are resistant to the development of autoimmune encephalomyelitis (22). When these mice are given leptin replacement, they become susceptible to autoimmune encephalomyelitis, suggesting that hypoleptinemia protects against an autoimmune response. In our patients with acquired lipodystrophy, hypoleptinemia does not seem to protect against other autoimmune diseases. Therefore, the relevance of the hypoleptinemic rodent model to humans is uncertain. Elucidating the effect of leptin on human immune function is further complicated by the observation that leptin replacement in patients with severe lipodystrophy causes a modest increase in T cell numbers in vivo (23). The same study also revealed that leptin therapy augments secretion of TNF-α from peripheral blood mononuclear cells in vitro but only to approximately normal levels of secretion.

Remarkably, patient NIH-9 had four different, perhaps autoimmune-mediated disorders, including AGL, T1D, MPGN type 1, and possibly autoimmune hepatitis. Whether drugs like pioglitazone or leptin may have exacerbated these disorders cannot be fully determined (11), but it is clear that there was no amelioration of these disorders when the drugs were discontinued.

Leptin replacement therapy has been shown to improve glycemic control, hypertriglyceridemia, and NASH in patients with generalized and partial lipodystrophy (3,4,5,10,24). In this paper, we demonstrate for the first time that leptin improved hyperglycemia and reduced triglyceride levels in two patients who have T1D superimposed on a background of AGL.