Insulin Resistance Syndrome in Subjects With Mutated RING Finger Protein TRIM37

The study group included 65 MUL patients (37 female) aged 1.1–55 years. Sixty-three were homozygous for the Finnish founder mutation (c.493–2A>G), and two were compound heterozygotes for the founder and a c.2212delG mutation. Both mutations predict a truncated TRIM37 protein (9,10). All but three of the MUL patients (95%) were born SGA, with a mean birth length and weight of 45 cm (−3.0 SD score [SDS]) and 2,300 g (−2.7 SDS), respectively. None presented catch-up growth, and the median final height was 137 cm (−4.7 SDS) for female and 152 cm (−3.6 SDS) for male subjects (Table 1). Six patients received growth hormone therapy at the time of blood sampling.

All patients underwent a physical examination with assessment of height, weight, and pubertal stage (according to the criteria of Tanner) (12) and ultrasound evaluation of the liver. Weight was presented as weight for height (percent deviation from the age-specific median weight for height of Finnish standards) (13), since it revealed the weight changes at different ages better than the BMI. The blood pressure was measured three times while the subjects were seated, and the last two measurements were averaged for analysis.

Blood glucose (venous whole-blood glucose) and serum insulin were measured after an overnight fast with concomitant measurement of serum leptin and plasma uric acid, alanine aminotransferase, aspartate aminotransferase, γ-glutamyl transferase, and triglycerides as well as plasma total, HDL, and LDL cholesterol. All but six young patients (aged from 1.1 to 5.0 years) underwent a 3-h oral glucose tolerance test (OGTT; glucose load 1.75g/kg, maximum 75 g) with measurements of blood glucose and serum insulin. Thirty of the patients had a second OGTT and laboratory follow-up 2–5 years after their first test. The OGTT was used to classify the subjects as having normal glucose tolerance, impaired glucose tolerance (IGT; the 2-h venous whole-blood glucose value from 6.7 to 9.9 mmol/l.), or type 2 diabetes (fasting venous whole-blood glucose ≥6.1 mmol/l or the 2-h value ≥10.0 mmol/l) on the basis of World Health Organization criteria (14). The degree of insulin sensitivity was determined from the ratio of fasting glucose and insulin levels (15) and the whole-body insulin sensitivity index was calculated from the OGTT values according to Stumvoll (16,17).

Clinical and laboratory parameters were correlated with age and weight for height, and the explanation rate was expressed as an R² value. No significant changes were noted in the pattern of distribution and in the R² value, when all glucose and insulin measurements in fasting (n = 95) or during the OGTT (n = 89) were analyzed, compared with the use of only one measurement form each patient (n = 65). The ethics committee of our hospital approved the study protocol, and all patients or their guardians gave informed consent.

Glucose metabolism in MUL patients showed a dramatic change with age. Patients aged <10 years had low fasting glucose and insulin levels, so that hypoglycemic values (blood glucose <3.4 mmol/l) were recorded in half (14 of 28) of the children (Fig. 1A and B, Table 1). Fasting glucose concentrations and the 2-h postload glucose concentrations during OGTT increased with advancing age. Among adults (aged >20 years), IGT or type 2 diabetes was diagnosed in 92% of the patients (IGT 42% and type 2 diabetes 50%), indicating serious deterioration in glucose homeostasis in young adult life. The median ages revealing IGT or type 2 diabetes were 18.3 (range 3.1–42.3) and 26.0 (9.5–55.2) years, respectively.

The changes in insulin levels were even more marked; abnormally high fasting (>15 mU/l) and postload peak (>150 mU/l) insulin values were recorded in 90% of the adults (24/26), and the peak insulin values ranged from 138 to 1,450 mU/l (median 430 mU/l) (Fig. 1A). Serum basal and postload insulin levels correlated well with the C-peptide levels, and the explanation rate (R² value) was 0.6121 and 0.657, respectively. The fasting glucose and insulin levels and Stumvoll indexes showed a 10- and 4-fold decrease with advancing age, indicating the development of severe insulin resistance (Fig. 1C and D, Table 1).

The development of insulin resistance was accompanied by changes in body composition and lipid metabolism. While prepubertal children were thin, clinically evident abdominal obesity started to develop after puberty, and 42% (14 of 34) of the adults were overweight (weight for height >20%) (Fig. 1E). The insulin levels and weight for height, however, correlated poorly, and only 28% (12/43) of the patients with a postload serum insulin level exceeding 150 mU/l were overweight. Serum leptin, total cholesterol, and triglycerides also increased with age (Fig. 1G and H, Table 1). The leptin levels correlated with the peak insulin levels (R² = 0.66) and weight for height (R² = 0.606). While growth hormone is known to temporarily increase fasting and postprandial insulin levels, the insulin levels of the six patients receiving growth hormone were not higher than the levels of other subjects in the same age category.

Ultrasonography of the liver showed increased echogenicity suggestive of fatty liver in all adolescents and adults and in 46% of the prepubertal children. Moreover, fatty liver was noted in all patients (n = 12) who underwent a liver biopsy (at the age of 1.8–35 years) and at autopsy in 85% (11/13) of historical MUL patients deceased at the age of 2.0–49 years. Atherogenic vascular changes were noted in two thirds of the autopsied patients. Ninety percent of the adults had elevated γ-glutamyl transferase levels (Table 1).

Acantosis nigricans was present in some children and was a frequent finding postpubertally (80%), particularly in axillae, groins, and the genital region. Serum uric acid was elevated in two thirds of the patients during the period of growth and remained elevated in nearly half of the adults (Table 1). Hypertension (>95th percentile for age and sex) was observed in 81% of the adults (Fig. 1F, Table 1), and, overall, 70% of the adults fulfilled the criteria for metabolic syndrome according to the National Cholesterol Education Program Adult Treatment Panel (18).

The results indicate that subjects with a rare monogenic disorder, MUL, develop severe IRS in early adulthood. The pathophysiology of this process remains to be solved. While it may be partly explained by poor intrauterine growth, other factors are most likely involved. High risk of IRS has been associated with SGA caused by an unfavorable mother-child environment followed by early postnatal weight gain to obesity (19,20). In MUL, however, the prenatal-onset growth failure is caused by a fetal gene defect, and only a modest weight gain is observed before puberty. Moreover, in our experience the link between SGA and IRS is not definitive, as subjects with Silver-Russell syndrome, who have a phenotype and growth pattern similar to MUL patients, do not show abnormalities in glucose metabolism, as studied by OGTTs (data not shown).

In most monogenic diseases leading to type 2 diabetes, the pathogenic process can be explained by dysfunction of the pancreatic β-cells or obesity (21). Contrary to this, these subjects presented exceptionally high postload serum insulin levels, and fatty liver as well as acanthosis nigricans were often already detectable in slim prepubertal MUL children. Additionally, the weight gain did not correlate with IRS. On the other hand, lipodystrophy with low leptin secretion (21) was not involved in the development of insulin resistance. In MUL, serum leptin levels were unexpectedly high compared with the weight gain. The highly elevated serum leptin levels together with early development of fatty liver seen in MUL suggest that accumulation of liver fat may be a crucial step in the development of IRS in these subjects.

The genetic defects in the Finnish MUL patients predict a nonfunctional truncated TRIM37 protein, a member of the TRIM subfamily of zinc finger proteins (9,10). While the physiological function of TRIM37 in vivo is not known, it is expressed in several tissues (22) and has been localized to peroxisomes in cell cultures (10). More recently, TRIM37 was found to possess TRIM domain–dependent E3 ubiquitin-ligase activity implying defective ubiquitin-dependent degradation of an as-yet-unidentified target protein in the pathogenesis of MUL (23). Interestingly, inherited variation in the peroxisome proliferator–activated receptor γ gene has been implicated in the pathogenesis of type 2 diabetes (24) and also with reduced fetal growth (4). Moreover, recent reports have demonstrated that members of the ubiquitin-proteosome pathway can affect the regulation of insulin signaling cascades and insulin action (6,25). In the light of these findings, the peroxisomal targeting and the functional link of TRIM37 to the ubiquitin pathway may provide novel clues to understanding the insulin action during the fetal period and the development of metabolic syndrome later in life. Thus, MUL could serve as an excellent model for both molecular and clinical studies on the development of type 2 diabetes and metabolic syndrome.

This study was supported by the Finnish Foundation for Pediatric Research, Finska Läkaresällskapet, Finnish State Grants TYH2308 and TYH3304, the Finnish Academy, and the Sigrid Juselius Foundation and Else and Wilhelm Stockmann Foundation.

J.K. is a fellow of the Helsinki Biomedical Graduate School.

We thank our patients, their families, and local physicians.