Metreleptin Improves Blood Glucose in Patients With Insulin Receptor Mutations

Abstract

Context:

Rabson-Mendenhall syndrome (RMS) is caused by mutations of the insulin receptor and results in extreme insulin resistance and dysglycemia. Hyperglycemia in RMS is very difficult to treat, and patients are at risk for early morbidity and mortality from complications of diabetes.

Objective:

Our objective was to study 1-year effects of recombinant human methionyl leptin (metreleptin) in 5 patients with RMS and 10-year effects in 2 of these patients.

Design and Setting:

We conducted an open-label nonrandomized study at the National Institutes of Health.

Patients:

Patients were adolescents with RMS and poorly controlled diabetes.

Intervention:

Two patients were treated with escalating doses (0.02 up to 0.22 mg/kg/d) of metreleptin for 10 years, including 3 cycles of metreleptin withdrawal and reinitiation. In all 5 patients, 1-year effects of metreleptin (0.22 mg/kg/d) were studied.

Outcome Measures:

Hemoglobin A1c (HbA1c) and body mass index (BMI) z-scores were evaluated every 6 months.

Results:

HbA1c decreased from 11.4% ± 1.1% at baseline to 9.3% ± 1.9% after 6 months and 9.7% ± 1.6% after 12 months of metreleptin (P = .007). In patients treated for 10 years, HbA1c declined with each cycle of metreleptin and rose with each withdrawal. BMI z-scores declined from −1.4 ± 1.8 at baseline, to −2.6 ± 1.6 after 12 months of metreleptin (P = .0006). Changes in BMI z-score correlated with changes in HbA1c (P < .0001).

Conclusions:

Metreleptin treatment for 12 months was associated with a 1.7% reduction in HbA1c; part of this improvement was likely mediated via decreased BMI. Metreleptin is a promising treatment option for RMS, but additional therapies are needed to achieve HbA1c targets.

Mutations of the insulin receptor result in extreme insulin resistance and dysglycemia in humans (1, 2). These disorders have a spectrum of clinical severity linked to the degree of residual activity of the insulin receptor (3). Three clinical syndromes have been described: infants with Donohue syndrome (also called leprechaunism) and children with Rabson-Mendenhall syndrome (RMS) have characteristic dysmorphic features and frequently have fasting hypoglycemia coupled with postprandial hyperglycemia (4), whereas patients classified as type A insulin resistance lack dysmorphisms (5). Most patients who survive beyond 2 years of age (RMS and type A insulin resistance) develop persistent hyperglycemia as endogenous insulin secretion declines (5, 6). Hyperglycemia in patients with insulin receptor mutations is extremely difficult to treat (7), and patients are at risk for early morbidity and mortality from microvascular complications of diabetes (5).

High doses of concentrated insulin have been used to treat patients with insulin receptor mutations in an effort to maximize any residual insulin receptor signaling. Recombinant IGF-1 has also been tried in small numbers of patients and appears to be most effective in those with the mildest form of the disease (8). In our experience, IGF-1 has not been useful in the most severely affected patients, and if its action is through residual insulin receptor activity, it is more logical and safer to use high-dose insulin. Other conventional glucose-lowering therapies, such as metformin, may be modestly beneficial in patients with insulin receptor mutations, but additional glucose-lowering treatments that do not require signaling through the insulin receptor are required for effective treatment of these conditions.

In 2004, our group reported the glucose-lowering effects of 10 months of recombinant human methionyl leptin (metreleptin) in 2 siblings with RMS (9). Leptin is an adipocyte-derived hormone and is a major regulator of energy metabolism in humans. Leptin signals through a cytokine receptor that bears no structural similarity or ligand-binding specificity with the insulin receptor. However, the postreceptor signal transduction cascades of the leptin and insulin receptors overlap at the level of phosphoinositide 3-kinase (PI3 kinase), and thus, there is theoretical reason to believe that pharmacologic treatment with leptin might increase postreceptor insulin signaling (10). In this study, we report the long-term (10 years) effects of metreleptin treatment in the initial 2 patients reported as well as short-term (1 year) effects of metreleptin in an additional 3 patients with RMS.

Patients and Methods

Patients were enrolled in an open-label study of metreleptin in individuals with low leptin levels (<12 ng/mL in females and <6 ng/mL in males) and a syndrome of extreme insulin resistance (NCT00085982). The study was approved by the Institutional Review Board of the National Institute of Diabetes and Digestive and Kidney Diseases. Patients and their parents provided written, informed consent/assent. Metreleptin was given as twice-daily sc injections. Metreleptin has a time to peak and half-life of approximately 3 hours and is renally cleared (11). The first 2 patients were siblings who initiated metreleptin at doses of 0.02 to 0.03 mg/kg/d; doses were titrated up over years to a maximum of 0.22 to 0.24 mg/kg/d. During the 10 years of treatment, metreleptin was withdrawn on 3 occasions for periods of 1 to 10 months and then restarted (Figure 1). These withdrawals were performed because the initial short-term treatment study concluded (first withdrawal) and to study the effect of metreleptin withdrawal on glycemia control (second and third withdrawals). In the 3 most recently enrolled patients, metreleptin was initiated at a dose of 0.22 to 0.24 mg/kg/d and continued at that dose without withdrawal. The dose of 0.22 to 0.24 mg/kg/d was intended to provide a supraphysiologic dose of leptin, equivalent to approximately 3 times endogenous leptin levels for individuals with normal body fat percentage, based on data provided by the manufacturer.

Figure 1.

Metreleptin doses (gray shaded area), HbA1c (blue triangles), and insulin dose (red inverted triangles) in 2 siblings with RMS treated with metreleptin for 10 years. Metreleptin was initiated at doses of 0.02 to 0.03 mg/kg/d, and gradually titrated up over 6 months to 0.06–0.09 mg/kg/d. After the 10-month pilot study ended, metreleptin was withdrawn for 10 months and subsequently restarted at doses of 0.06 to 0.09 mg/kg/d for 18 months. After a second withdrawal period of 1 to 3 months, these patients restarted metreleptin, and the doses were gradually increased over 2.5 years to a maximum of 0.22 to 0.24 mg/kg/d. After 5 years of continuous metreleptin treatment, a third withdrawal of 2.5 months was performed, followed by reinitiation of metreleptin at the previous doses and another 2 years of follow-up to date. Periods of metreleptin initiation or dose increase were generally associated with reductions in HbA1c, whereas HbA1c rose during periods of metreleptin withdrawal (black arrows).

Patients underwent metabolic testing at the National Institutes of Health approximately every 6 months. Fasting laboratory parameters were measured using standard clinical methods and included hemoglobin A1c (HbA1c), IGF-1, GH, testosterone, total cholesterol, triglycerides, high-density lipoprotein (HDL), and low-density lipoprotein. Leptin was measured by RIA using a commercial kit (Linco Research) approximately 12 hours after the last metreleptin dose; this assay does not distinguish between endogenous leptin and exogenous metreleptin. Oral glucose tolerance testing (OGTT) was performed after an overnight fast with 1.75 g/kg (maximum 75g) dextrose oral solution, with blood samples obtained at −10, 0, 30, 60, 90, 120, 150, and 180 minutes for glucose, insulin, C-peptide, and glucagon; area under the curve (AUC) was calculated using the trapezoidal method. For patients treated with insulin, the last dose of insulin (U-500 regular human insulin) was given the day before the OGTT. Body fat was determined using a dual-energy x-ray absorptiometer (model QDR 4500; Hologic).

Statistical analysis

Effects of metreleptin on HbA1c and body mass index (BMI) z-scores were tested using a 12-month period during which all patients were treated with comparable doses of metreleptin (0.22–0.24 mg/kg/d) and were on stable or slightly decreasing doses of other diabetes medications, including insulin. For subjects RMS-1 and RMS-2, this was the year of metreleptin treatment after the third withdrawal; for other subjects, this was the first year of leptin treatment. The z-scores for height, weight, and BMI in subjects under 20 years of age were calculated based on Centers for Disease Control growth data for the US population. Because the cohort included subjects both younger and older than 20 years, BMI z-scores for subjects older than age 20 were calculated using normative data for 20-year-olds. Changes in HbA1c and BMI z-scores over time were tested using mixed models (PROC MIXED). Post hoc Tukey correction was used to correct for multiple comparisons at 0, 6, and 12 months of leptin treatment. Statistical analyses were conducted using SAS version 9.3. Data in the text are reported as mean ± SD, and graphs show mean ± SEM except as noted.

Results

Patient characteristics

Baseline subject characteristics are shown in Table 1. All patients had homozygous or compound heterozygous mutations of the insulin receptor, and a phenotype consistent with RMS, including extreme insulin resistance (severe acanthosis, acrochordons, and hyperinsulinemia with high insulin to C-peptide ratios), dental anomalies (early secondary dentition, large teeth, dental crowding, and malocclusion), and short stature. All patients had nephrocalcinosis and/or nephrolithiasis noted at baseline or during follow-up. Four patients had frequent otitis media, and 3 had cholesteatoma.

Table 1.

Baseline Characteristics of Patients With RMS Due to Mutations in the Insulin Receptor

	Subject ID
	RMS-1a	RMS-2a	RMS-5	RMS-6	RMS-7
Age, y	13	11	12	15	12
Sex	Male	Female	Female	Male	Male
Race	Asian	Asian	Black	Caucasian	Asian
Genetic mutation	Pro193Leu	Pro193Leu	Ile119Met	Ser635Leu	Tyr30X (40)
	Pro193Leu	Pro193Leu	Ile119Met	Del exons 9 and 10	Glu238Lys
HbA1c, %	9.6	9.2	10.1	11.9	10.6
Insulin dose, U/d	300	0	1450	0	0
Insulin, μU/mLb	327	219	114	256	300
C-peptide, ng/mLb	3.8	6.0	0.1	6.4	6.9
Glucose, mg/dL	226	151	54	168	160
Puberty	G III, PH III, testes 8–10 mL	B III, PH III, clitoromegaly	B III, PH I, clitoromegaly (at age 13)	G III, PH II, testes 15 mL	G III, PH III, testes 15 mL
Height z-score	−4.46	−2.57	−1.29	−3.81	−1.26
Weight z-score	−5.20	−3.76	−1.07	−2.14	−1.02
BMI z-score	−2.40	−2.83	−0.59	0.53	−0.29
IGF-1, ng/mLc	24	28	<25	38	90
Total cholesterol, mg/dL	135	138	119	206	160
Triglycerides, mg/dL	42	33	55	56	51
HDL, mg/dL	70	80	85	85	88
LDL, mg/dL	72	70	23	110	62
Leptin, ng/mL	3.42	4.46	8.20	1.69	2.81
Duration of metreleptin treatment, mod	128	128	17	12	14

Abbreviations: B, breast stage; G, genital stage; LDL, low-density lipoprotein; PH, pubic hair stage.

^aBaseline characteristics for subjects 1 and 2 are given for the time of first initiation of metreleptin treatment.

^bThe last dose of U-500 regular human insulin was given the day before data collection in subjects RMS-1 and RMS-5. High insulin to C-peptide ratios are characteristic of patients with insulin receptor mutations due to impaired insulin clearance secondary to the dysfunctional insulin receptor.

^cNormal range for Tanner stage III boys is 94–765. Normal range for Tanner stage III girls is 145–760.

^dIn subjects 1 and 2, time elapsed since first course of metreleptin was initiated.

All patients had poorly controlled diabetes (HbA1c >9%) at the time of metreleptin initiation. All had entered puberty, as indicated by breast Tanner staging (in girls) or testicular volume (in boys). In the 2 female patients, hyperandrogenism was present, with clitoromegaly and peak testosterone levels of 701 (RMS-2) and 432 (RMS-5) ng/dL. Subject RMS-2 ultimately underwent oophorectomy at age 19 years for management of hyperandrogenism. IGF-1 levels were low in all patients, whereas GH levels were normal. Lipid levels were typical for patients with insulin receptor mutations, with low-normal triglycerides and high-normal HDL levels. In RMS-1 and RMS-2, glucagon levels were obtained during OGTT during a time when they were not treated with metreleptin. In both cases, glucagon appropriately suppressed after the glucose load (from 35 to 17 pg/mL in RMS-1 and from 71 to 28 pg/mL in RMS-2) and remained below fasting levels for 180 minutes after glucose ingestion.

Diabetes management

All patients were treated with metformin throughout the study, and one (RMS-6) was treated with pioglitazone. Two patients (RMS-1 and RMS-5) were treated with U-500 regular human insulin at baseline and throughout follow-up; one patient (RMS-2) started U-500 insulin 1 year after starting metreleptin and continued to take insulin for the next 9 years. Two patients (RMS-6 and RMS-7) were never treated with insulin.

Effects of 12 months of high-dose metreleptin

Leptin levels rose from 3.9 ± 2.5 ng/mL before metreleptin treatment to 28.4 ± 15.9 (range, 4.5–46.2) after 12 months of high-dose metreleptin.

HbA1c decreased from 11.4% ± 1.1% before metreleptin to 9.3% ± 1.9% after 6 months and 9.7% ± 1.6% after 12 months of metreleptin treatment (Figure 2A; P = .007 for overall treatment effect, P = .008 for 0 vs 6 months, and P = .027 for 0 vs 12 months). Similarly, glucose AUC during the OGTT declined from 62 485 ± 9748 mg/dL × 180 minutes before metreleptin to 56 039 ± 14 642 after 6 months and 48 839 ± 15 779 after 12 months (Figure 2B; P = .01 for overall treatment effect, P = .05 for 0 vs 6 months, and P = .01 for 0 vs 12 months). Insulin and C-peptide AUC did not change significantly during the 12 months of metreleptin treatment (Figure 2B; P = .5 and P = .77 for insulin and C-peptide, respectively).

Figure 2.

A, Effects of 1 year of high-dose metreleptin in 5 patients with RMS. HbA1c decreased from 11.4% to 9.7% (P = .007). B, OGTTs showed a significant decline in glucose AUC (black circles with solid line, P = .01) and no change in insulin (black squares with dashed line) or C-peptide (open triangles with dotted line) AUC. C and D, BMI z-scores (C) decreased from −1.4 to −2.5 (P = .0006), and percent body fat (D) declined from 20.5% to 16.1% (P = .14).

BMI z-scores declined with metreleptin treatment, from −1.4 ± 1.8 before metreleptin to −2.5 ± 1.6 after 6 months and −2.6 ± 1.6 after 12 months of metreleptin treatment (Figure 2C; P = .0006 for overall treatment effect, P = .0018 for 0 vs 6 months, and P = .0009 for 0 vs 12 months). After adjustment for changes in BMI z-score, the change in HbA1c with metreleptin was of only borderline statistical significance (P = .069). Body fat percentage declined nonsignificantly from 20.5% ± 5.1% before metreleptin to 16.1% ± 6.5% after 12 months (Figure 2D; P = .14).

Fasting glucagon levels were obtained before metreleptin in 4 subjects (36, 71, 32, and 30 pg/mL in RMS-1, RMS-2, RMS-6, and RMS-7, respectively; normal, <80 pg/mL). There was no consistent trend in glucagon after 6 months of metreleptin treatment in 3 subjects (49, 42, and 32 pg/mL in RMS-1, RMS-2, and RMS-6, respectively). Triglyceride levels remained low during 1 year of metreleptin (60 ± 34 mg/dL at baseline vs 58 ± 22 mg/dL at 12 months, P = .9), and HDL levels remained high (76 ± 15 mg/dL at baseline vs 72 ± 15 mg/dL at 12 months, P = .26). IGF-1 levels remained low during treatment (46 ± 28 ng/mL at baseline vs 46 ± 32 ng/mL at 12 months, P = .9).

Long-term effects of metreleptin

In RMS-1 and RMS-2, periods of metreleptin initiation or dose increase were generally associated with declines in HbA1c, whereas periods of metreleptin withdrawal were associated with increases in HbA1c (Figure 1). RMS-1 was treated with concentrated U-500 insulin at doses of approximately 300 U/d throughout the 10-year period, whereas subject RMS-2 required progressively higher doses of U-500 insulin after the first 10 months of the study, gradually increasing to 1100 U/d by month 92. However, changes in HbA1c in RMS-2 temporally correlated much better with metreleptin initiation or withdrawal than with increases in insulin dose.

Among all 5 subjects throughout the entire duration of metreleptin therapy, short-term incremental (from one data collection to the next) changes in BMI z-score were linearly associated with changes in HbA1c (P < .0001, R² = 0.26).

Other clinical features of RMS

Female subjects developed severe hyperandrogenism during puberty. RMS-2 had a peak testosterone level of 701 ng/dL (normal adult female range, <80 ng/dL) at age 19 and underwent bilateral oophorectomy to control hyperandrogenemia at age 20. RMS-5 had a peak testosterone level of 432 ng/dL at age 13 and underwent a trial of hormonal contraceptives in an unsuccessful attempt to control hyperandrogenism.

All subjects had short stature of variable severity. RMS-6 and RMS-7 were treated with GH or IGF-1 before starting metreleptin without improvement in growth velocity or glycemia. RMS-1 attained an adult height of 146 cm (−4.2 SD). RMS-2 attained an adult height of 148 cm (−2.4 SD). The remaining subjects were still growing at the time of last data collection and had heights of 0.8, 3.7, and 0.7 SD below the means for age and sex. Other than poorly controlled diabetes, no subject had an additional medical illness contributing to growth failure (eg, impaired thyroid, renal, or hepatic function or chronic inflammatory disease).

All subjects maintained normal lipid parameters, and 4 of 5 subjects underwent ultrasonography of the liver with no evidence of steatosis.

Discussion

In this single-arm, open-label study, we demonstrated statistically significant declines in HbA1c over 1 year of metreleptin therapy in 5 patients with RMS. The decline in HbA1c was maximal after 6 months of treatment, and HbA1c tended to rise thereafter. Despite metreleptin treatment, HbA1c remained markedly elevated between 9% and 10%, and only 1 subject achieved an HbA1c in the target range of <7% at a single time point (RMS-7, after 6 months of metreleptin). Despite this incomplete response, the observed reduction in HbA1c from 11.4% to 9.7% should substantially reduce the risk of microvascular complications of diabetes; for example, data from the Diabetes Control and Complications Trial suggest that this reduction in HbA1c should correspond to a 60% reduction in risk of retinopathy (12). Nonetheless, additional treatments are needed to further reduce HbA1c and the risk of microvascular complications of diabetes.

Metreleptin withdrawal was repeatedly associated with a marked rise in HbA1c in 2 subjects who were treated with metreleptin over the course of 10 years, suggesting that this medication does have a sustained effect to reduce blood glucose in this patient population. Because the natural history of HbA1c change over time in patients with RMS has not been well established, it is possible that the rise in HbA1c over time during continued metreleptin treatment represents an amelioration of a larger rise in HbA1c that might occur in the absence of metreleptin. Alternatively, the rise in HbA1c seen with long-term metreleptin may reflect decreased compliance with a twice-daily injectable medication during adolescence, a time when diabetes control becomes notoriously difficult. Acquired resistance to metreleptin is an unlikely explanation for the waning benefit of metreleptin over time in RMS, because metreleptin treatment in patients with lipodystrophy results in sustained decreases in HbA1c over many years (13), and weight loss in patients with congenital leptin deficiency is also sustained over the long term (14). Randomized clinical trials are needed to establish the true long-term effect of metreleptin on glycemia in patients with RMS.

A weakness of this study was the inability to fully elucidate the mechanisms by which metreleptin improved glucose in RMS. Although we initially hypothesized that metreleptin might ameliorate hyperglycemia in RMS via cross talk between leptin and insulin postreceptor signaling, we could not test this hypothesis directly in the current study. Another likely explanation for the improvement in glycemia seen after metreleptin treatment is metreleptin's effect on energy intake. Leptin is a major signal of satiety; leptin acts in the arcuate nucleus of the hypothalamus to suppress orexigenic pathways via agouti-related peptide (AgRP)/neuropeptide Y (NPY) cells and to stimulate anorexigenic pathways via pro-opiomelanocortin (POMC)/cocaine-amphetamine–associated transcript (CART) neurons (15). Subcutaneous administration of metreleptin in conditions of low endogenous leptin levels (congenital leptin deficiency and lipodystrophy) decreases food intake and energy storage (14, 16), and alters central nervous system activity in regions associated with hunger and satiety (17). Reduction in acute energy intake is well known to reduce glucose levels in both type 1 and type 2 diabetes (18–23). We did not directly measure food intake in our patients before or during metreleptin treatment; however, we used changes in body weight (measured as change in BMI z-scores) as a proxy measure for food intake. We found that changes in body weight explained 26% of the variance in HbA1c change, supporting the hypothesis that metreleptin's effect on food intake is a likely mediator of at least part of its effect on glycemia in RMS.

Rodent models with genetic knockout of the insulin receptor have been used to improve understanding of the pathophysiology of insulin resistance. In contrast to humans with Donohue syndrome or RMS, total-body knockout of the insulin receptor in rodents causes neonatal lethality from diabetic ketoacidosis (24). The liver-specific insulin receptor knockout, or LIRKO, mouse is the most similar to RMS, with hyperinsulinemia and mild hyperglycemia in lean animals (24). Unlike humans with RMS, LIRKO mice have high circulating leptin levels but maintain normal leptin responsiveness (25, 26). Pharmacologic leptin treatment in these animals causes substantial reduction in food intake (by 40%), supporting the hypothesis that reduced food intake may play a major role in the improved glycemia seen after leptin treatment (26).

In patients with lipodystrophy, improvements in glycemia seen with metreleptin treatment are thought to be mediated via leptin-induced improvements in insulin sensitivity (27, 28). This improvement in insulin sensitivity may, in turn, be mediated via reductions in ectopic lipid accumulation (and hence, lipotoxicity) in the liver and muscle (27), and these metabolic benefits are maintained over a prolonged timeframe (13). These explanations are unlikely to play a substantial role in patients with insulin receptor mutations because such patients have major static genetic defects in insulin sensitivity and they have normal lipid profiles and do not have ectopic lipid accumulation in the liver (29). We could not assess insulin sensitivity in this study because fasting and OGTT-derived indices of insulin resistance are not valid in extreme insulin resistance and the gold-standard hyperinsulinemic-euglycemic clamp cannot be effectively performed because insulin doses as high as 500 mU/m²/min fail to elicit any significant insulin-mediated glucose disposal (30). Changes in β-cell function were unlikely to account for the improvement in HbA1c because insulin AUC during the OGTT did not change with metreleptin treatment. Metreleptin might have improved glycemia by reducing endogenous glucose production, but this was not measured in the current study. Recombinant IGF-1 has been tried in patients with insulin receptor mutations, based on the known hypoglycemic action of IGF-1 due to substantial homology between the insulin and IGF receptors. As expected, metreleptin treatment had no effect on the low IGF-1 levels seen in our patients, and thus any benefit of metreleptin on glycemia was not mediated via IGF-1.

Leptin has been proposed to improve glycemia in rodent models of type 1 diabetes by reducing glucagon secretion (31). Type 1 diabetes is characterized by fasting glucagon excess, with paradoxical increased glucagon secretion after a mixed meal (32) or glucose load (33), and hence suppression of glucagon can logically be expected to aid in blood glucose control. Although limited data are available, patients with RMS, like healthy individuals (33), do not have elevated fasting glucagon levels, have sustained suppression of glucagon below fasting levels after oral glucose, and do not have suppression of glucagon secretion with metreleptin treatment, making this an unlikely mechanism to account for improved glycemia in these patients.

Although the focus of this study was glycemia control, we observed other features of RMS during follow-up. As previously reported in disorders of severe insulin resistance (2, 5, 34, 35), the two female patients had severe hyperandrogenism due to pathologic stimulation of ovarian steroidogenesis by high circulating insulin. Patients of both sexes had short stature, a feature of both Donohue syndrome and RMS that is thought to be due to the severe defect in insulin signaling (3, 5, 36). All patients had nephrocalcinosis; although this is known to be associated with RMS, the pathophysiology underlying this disorder is not clear (34).

In conclusion, metreleptin treatment improved HbA1c after 1 year in 5 patients with RMS and continued to show some benefit after 10 years in 2 patients. This improvement in glycemia control was likely mediated in part by reductions in food intake, but more studies are needed to better understand the mechanism of action of metreleptin in these patients. The beneficial effects of metreleptin on HbA1c appeared to wane over time, and metreleptin treatment, with or without insulin, was insufficient to achieve target HbA1c levels of less than 7% in most patients. Additional, insulin receptor-independent treatments are needed to reduce complications of diabetes and mortality in patients with insulin receptor mutations. Such therapy might relate to expansion of non–insulin-mediated glucose-utilizing tissues, such as brown fat (30). Molecules such as irisin (37), FGF-21 (38), or environmental temperature modulation (39) may be useful in this regard.