Risk Factors for Diabetes in Familial Partial Lipodystrophy, Dunnigan Variety Free

All subjects gave written informed consent. Appropriate institutional review boards approved the protocol. Only adults (>18 years of age) were included in this analysis. FPLD was diagnosed on the basis of characteristic phenotype and the presence of missense mutations in the LMNA gene. All unaffected relatives were genotyped for the presence of the missense mutation identified in their family and had only wild-type alleles. We studied 24 men and 52 women with FPLD belonging to 18 different families. All families were of European origin, except family F2600, which was of Asian Indian origin. Most of these pedigrees have been previously reported (7,9,11,12). All the families were evaluated at the University of Texas Southwestern Medical Center at Dallas, except F1000 and F2500 families (24 subjects), who were evaluated at the National Institutes of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, Maryland. Nineteen of these subjects (six women and three men with diabetes and five women and five men without diabetes) visited the General Clinical Research Center of the University of Texas Southwestern Medical Center at Dallas for evaluation; information and blood samples on the other 33 patients were obtained by mail as described previously (10).

All subjects reported demographic data, date of birth, height, and weight, as well as health history, particularly related to the presence or absence of diabetes, dyslipidemia, hypertension, acanthosis nigricans, parity, menopausal status, menstrual irregularity, hirsutism, and polycystic ovaries and use of medications. Diabetes was diagnosed on the basis of previous history, use of hypoglycemic medications, or if fasting serum glucose concentration exceeded 7.0 mmol/l. Hypertension was diagnosed on self-report or use of antihypertensive medications or if either systolic or diastolic blood pressure exceeded 140 or 90 mmHg, respectively.

In the 19 subjects evaluated at Dallas, skinfold thickness was measured with a Lange caliper (Cambridge Scientific Industries, Cambridge, MD) at the chin and five truncal (chest, mid-axillary, abdomen, subscapular, and suprailiac) and six peripheral (biceps, triceps, forearm, hip, thigh, and calf) sites on the right side of the body. The mean of three repeat measurements at each site was calculated.

Blood was collected after a 12-h overnight fast for analysis of HbA_1c and for serum chemistry profile, lipoproteins, insulin, and leptin concentrations. Fasting serum samples were analyzed for cholesterol, triglycerides, and HDL cholesterol as described previously (10). Blood HbA_1c was measured using ion exchange high-performance liquid chromatography (Bio-Rad Laboratories, Hercules, CA). Serum insulin and leptin levels were determined by radioimmunoassay using commercial kits (Linco Research, St. Charles, MO).

Mutational analysis of LMNA was performed on all the patients by direct sequencing of the coding region and the splice-site junctions, as described previously (7).

Categorical data were compared by Fisher’s exact test. The biochemical parameters and anthropometric data were compared using Wilcoxon’s rank-sum test. Data are expressed as median (range), and P value <0.05 was considered statistically significant.

All patients, except one man, were previously diagnosed to have diabetes. Forty-three percent of them were treated with insulin, 23% with oral hypoglycemic agents, and 34% with nonpharmacological measures, including diet and physical activity. The prevalence of diabetes among women was approximately three times higher than in men with FPLD (56% vs. 17%, respectively; P < 0.001). Although comparative data on genotypes, clinical characteristics, and biochemical parameters in both men and women with and without diabetes are presented, main conclusions are limited to the data on women because of small sample size of men with diabetes. Median age of onset of diabetes was 30 years (range, 17–51 years) in women and 34 years (25–39 years) in men. Whereas the prevalence of hypertension was similar in women and men (42% vs. 26%; P = 0.3), more women reported taking lipid-lowering medications (92% vs. 55%; P = 0.001).

Comparison of women with and without diabetes revealed similar age and prevalences of type 2 diabetes in unaffected relatives, acanthosis nigricans, polycystic ovarian syndrome, and premenopausal status (Table 1). Versus women without diabetes, however, those with diabetes had significantly higher BMI (P = 0.03), increased use of lipid-lowering medications (P = 0.002), and lower rates of nulliparity (P = 0.04) (Table 1). The prevalence of hypertension tended to be slightly higher among women with diabetes, although it did not reach statistical significance (P = 0.08).

Comparison of men with and without diabetes revealed similar age, BMI, history of type 2 diabetes in unaffected relatives, hypertension, and acanthosis nigricans. Two of the four men with diabetes were receiving lipid-lowering medications, versus none among men without diabetes (P = 0.02).

Skinfold measurements were available in six women with diabetes and five women without diabetes, and in three men with diabetes and six men without diabetes. Versus women without diabetes, those with diabetes had significantly increased skinfold thickness at the chin (10.4 mm [7.3–14.7 mm] vs. 20.3 mm [17–33 mm], respectively; P = 0.01) and at the biceps (4.5 mm [3–5.7 mm] vs. 6 mm [4–8 mm]; P = 0.03) (Fig. 1A). There were no significant differences in the skinfold thickness of the other peripheral and truncal sites between women with and without diabetes (Fig. 1A). In comparison with normal values (13), all women with FPLD had markedly reduced peripheral skinfold thickness at the triceps and thigh regions; however, truncal skinfolds at the chest, axillary, subscapular, abdominal and suprailiac regions ranged from <10th percentile to >90th percentile. No statistically significant differences were seen in the skinfold measurements between the FPLD men with and without diabetes (Fig. 1B). All men tended to have reduced peripheral skinfold thickness compared with normal values (14).

Plasma glucose and blood HbA_1c concentrations were higher in subjects with diabetes; however, plasma insulin levels were not significantly different (Table 2). Fasting plasma triglycerides and uric acid levels were higher in women with diabetes versus those without diabetes (P = 0.002 and 0.04, respectively), but plasma total cholesterol and HDL cholesterol levels were similar. Serum leptin levels were similar in women with and without diabetes (3.4 and 3.6 ng/ml; P = 0.9). There were no differences in plasma triglycerides, cholesterol, and HDL cholesterol concentrations between the two groups of men. The four male subjects with diabetes had lower serum leptin levels than the men without diabetes (median values, 0.9 vs. 2.9 ng/ml; P = 0.03).

We found eight different missense mutations in the LMNA gene in these patients (Table 3). Some of these mutations have been previously described by us (7,11,12); R419C and L515E are novel mutations, involving exons 7 and 9. All previous mutations in FPLD have been reported in exons 1, 8, and 11. Overall results revealed that the genotype did not predict increased prevalence of diabetes, although women carrying R482Q tended to have somewhat higher prevalence of diabetes than those with R482W (58 vs. 38%; P = 0.3).

Our data suggest that women with FPLD may be predisposed to develop diabetes owing to excess fat deposition in areas not affected by lipodystrophy. This was supported by the finding of higher BMI in women with diabetes versus those without diabetes. Since the loss of fat from most of the peripheral and truncal subcutaneous sites (except biceps) was similar between women with and without diabetes, this difference in BMI may relate to excess fat deposition in the head, neck, and intraabdominal regions. Certainly, women with diabetes had excess chin (submental) fat versus those without diabetes. Increased chin fat may represent deposition of fat in the areas unaffected by lipodystrophy such as intra-abdominal depots, and as such may be accompanied by fat deposition in other important organs such as the liver and skeletal muscles. Interestingly, patients with type 2 diabetes have been reported to have increased subscapular, chest, and abdominal skinfold thickness and decreased thigh and calf skinfold thickness (15,16).

Another predisposing factor for diabetes in women was parity, as less nulliparous women had diabetes. Kritz-Silverstein et al. (17) reported that parity was an independent risk factor for future development of type 2 diabetes, but other studies have found that the magnitude of this effect is reduced after adjusting for age, socioeconomic status, and BMI (18). Pregnancy is a well-known diabetogenic state and is a time of efficient fat storage. Because pregnant women with FPLD may not be able to accumulate fat in most of the subcutaneous regions, which are affected by lipodystrophy, they may accumulate excess fat in the face, neck, and intraabdominal region, predisposing them to diabetes.

Previous studies have identified dyslipidemia, particularly hypertriglyceridemia, as a predisposing risk factor for type 2 diabetes (19,20). Consistent with those studies, hypertriglyceridemia was also associated with predisposition to diabetes among women with FPLD. However, this association may be due to poor glycemic control in FPLD women with diabetes. Of note, a previous study reported higher serum triglyceride concentrations in FPLD patients with diabetes even when glycemic control was excellent (21). The onset of hypertriglyceridemia has reportedly antedated the onset of hyperglycemia in FPLD patients (21,22). Women with diabetes also had higher serum uric acid levels than those without diabetes. Interestingly, fasting serum insulin levels were not significantly different among FPLD women with and without diabetes. However, only nine women with diabetes had fasting serum insulin concentrations available because those on hypoglycemic therapy were excluded from this analysis.

Our anthropometric data, although limited, indicate that severity of lipodystrophy does not differ among FPLD women with and without diabetes, as most of the peripheral and truncal skinfold thicknesses were similarly reduced in the two groups. Serum leptin levels also were similarly reduced in the women with and without diabetes, compared with normal women matched for BMI (21). We have recently reported that serum adiponectin levels are reduced in FPLD subjects with diabetes versus those without diabetes (4.3 [1.9–12.7] vs. 7.6 [2.7–23.2] μg/ml, respectively; P < 0.05) (22). Reduced plasma adiponectin levels have been reported in other insulin-resistant states as well (23,24). These data suggest that FPLD patients with diabetes may be more insulin resistant than those without diabetes.

Another important aim of the study was to identify if any particular LMNA variant carried a high risk of diabetes among FPLD subjects. On the basis of study of a small pedigree, we have previously reported that R582H mutation may cause mild lipodystrophy as well as less severe metabolic complications, possibly because this particular mutation in exon 11 of LMNA affects transcription of lamin A only, whereas other previously reported mutations in FPLD patients in exon 1 and 8 affect transcription of both lamins A and C (11). However, our study did not have enough patients with some of the rare LMNA mutations to perform meaningful comparisons. When the two most prevalent mutations among FPLD subjects (R482Q and R482W) were compared, women carrying R482Q had somewhat higher prevalence of diabetes than those carrying R482W; however, this difference did not achieve statistical significance. Therefore, overall, prevalence of diabetes was not related to specific LMNA mutations. However, it is likely that insulin resistance in some of the FPLD patients may be due to modifier alleles of other genes implicated in causing insulin resistance.

Interestingly, some of the well-recognized risk factors for development of type 2 diabetes, such as, age, family history of type 2 diabetes, hypertension, and acanthosis nigricans (25,26), were not predictive of diabetes in women with FPLD. It is likely that the LMNA mutation may impart such a high risk of developing diabetes that it overshadows potential influence of these established risk factors for type 2 diabetes. FPLD being a rare disease, we consider data on 76 individuals as a large collection of subjects. Nonetheless, compared with previous population-based studies in type 2 diabetes (26,27), the number of subjects in our study is certainly limited and thus, the study may have lacked power to identify some of these risk factors. The lack of power certainly affected comparison of FPLD men with and without diabetes, since only four men had diabetes. Our study did not assess dietary intake and physical activity, which may be important confounding variables. Further, our data were cross-sectional and thus confirmation of our findings will require prospective longitudinal studies.

We conclude that women with FPLD are more predisposed to diabetes than men. This tendency among women may be due to increased adiposity in regions not affected by lipodystrophy, as reflected by increased submental deposition of subcutaneous fat. Furthermore, parity may predispose women with FPLD to diabetes. Age, menopausal status, family history of type 2 diabetes, and LMNA variants do not predict predisposition to diabetes.

The study was supported in part by the National Institute of Health Grants M01-RR00633 and R01-DK54387 and by the Southwestern Medical Foundation.

We are grateful to the patients and their families for participating in the studies; Angela Osborn for technical assistance and illustrations; Beverley Adams Huet for statistical analysis; Drs. David C. Robbins, Robert Kreisberg, Mark D. Shepherd, Andrea Dunaif, David Feinstein, Margo Denke, Irene Sills, Noralane M. Lindor, Tu T. Nguyen, Evelyn Cintron, and Stephen Aronoff for patient referral; and the staff at the General Clinical Research Center for patient care and support.