LMNA mRNA Expression Is Altered in Human Obesity and Type 2 Diabetes
Objective: The aim of this study was to analyze lamin A/C mRNA levels in abdominal subcutaneous adipose tissue in two conditions—obesity and type 2 diabetes—that share common inflammatory and metabolic features, and to assess their relationship with selected inflammatory and adipogenic genes.
Methods and Procedures: This is a cross-sectional study involving 52 nondiabetic and 54 type 2 diabetes patients. Anthropometrical and analytical measurements (glycemic, lipidic, and inflammatory profiles) were performed, and mRNA expression was determined using real-time PCR.
Results: Lamin A and C isoforms are expressed differentially. Lamin A/C mRNA levels were increased in obese and in type 2 diabetes patients. We also observed a strong relationship between both isoforms (B = 2.218, P < 0.001) and among lamin C mRNA expression and adipogenic (sterol-responsive element binding protein-1 (SREBP1c)) and inflammatory (interleukin-6 (IL-6)) markers (B = 0.854, P = 0.001, and B = 0.557, P < 0.001, respectively).
Discussion: These data suggest that lamin A/C may be involved in the adipocyte gene profile observed in obesity and type 2 diabetes.
Lamins are nuclear proteins that polymerize to form the nuclear lamina, a filamentous network located just below the inner nuclear membrane that assists in maintaining nuclear membrane integrity and in controlling gene expression (1,2). A-type lamins are the product of the LMNA gene, which contains 12 exons. By alternative splicing of exon 10, LMNA encodes lamins A and C (lamin A/C) as major products (3).
Many different mutations in the LMNA gene cause laminopathies, which result in different syndromes, as well as others with overlapping features. The precise mechanism by which defective A-type lamins cause these various syndromes remains unclear, but this could be an argument for common mechanisms in different conditions such as aging, regional adiposity, type 2 diabetes, and cardiac and skeletal muscle disorders (4).
Interestingly, the LMNA gene is located in a chromosomic region (1q21.2-21.3) (ref. 5), which has shown linkage with type 2 diabetes in several populations. Additionally, several association studies have been carried out, most of which concentrated on a coding variant in exon 10 (rs4641). Despite inconsistent results they showed an association with obesity-related traits (6,7,8,9,10,11,12). Recently, in two studies with larger cohorts, this polymorphism has been modestly associated with type 2 diabetes susceptibility (13,14). Despite all these data, LMNA expression was not analyzed in the subcutaneous adipose tissue of these patients.
A-type lamins have been reported to interact with the transcription factors sterol-responsive element binding protein-1 (SREBP1), c-Fos, MOK2, BAF, and GCL, and with the retinoblastoma gene product (pRb) (15,16,17,18,19,20,21,22,23), some of which play a key role in adipocyte differentiation (SREBP1, pRb). Furthermore, the interaction of A/C lamin proteins with other nuclear lamina-interacting proteins and with various transcription factors may play an important role in regulating cell differentiation (24,25).
Mutant A/C lamins may disrupt interaction with chromatin or other nuclear proteins, resulting in apoptosis of the adipocytes. In this way, some FPLD (Dunnigan-type familial partial lipodystrophy) mutations exhibit in vitro-reduced lamin A binding to SREBP1. Likewise, N195K and R386K LMNA mutations cause formation of nuclear lamin A and prelamin A aggregates that recruit proteins such as pRb and SREBP1a (26) that are critical in adipogenesis.
As far as mRNA is concerned, similar levels in expression of lamin A and C isoforms in both preadipocytes and adipocytes have been previously described at both mRNA and protein levels (27). However, an altered ratio between these isoforms has been described in a lung carcinoma cell line (28). Thus, changes in the balanced lamin A/C ratio could reflect a pathogenic condition.
All these findings led us to consider the possibility that lamin A/C could be involved in the adipose tissue biology of obesity and type 2 diabetes, two diseases in which adipose tissue plays an important role in their pathogenesis. These two conditions share common metabolic features, such as insulin resistance and a systemic and local (adipose tissue) subtle chronic inflammatory environment. The aim of this study was to analyze the lamin A/C mRNA levels in abdominal subcutaneous adipose tissue in these conditions and to assess its relationship with adipogenic and inflammatory markers.
Methods and Procedures
Obesity and type 2 diabetes cohort. A cohort of 106 subjects was recruited at the Hospital Universitari Joan XXIII (Tarragona, Spain). All were of white origin and reported that their body weight had been stable for at least 3 months before the study. They had no systemic disease other than obesity or type 2 diabetes, and all had been free of any infections in the month before the study. Liver and renal diseases were specifically excluded by biochemical work-up.
Obesity cohort. Subjects were classified by BMI according to the World Health Organization criteria (29). Using these criteria, there were 15 nonobese, 13 overweight, 10 nonseverely obese, and 14 severely obese nondiabetic subjects (Table 1).
Table 1. Obesity cohort
Type 2 diabetes cohort. Patients were classified as having type 2 diabetes according to the American Diabetes Association criteria (30). Variability in metabolic control was assessed by stable glycated hemoglobin A1c (HbA1c) values during the previous 6 months. Gathering these criteria, there were 54 type 2 diabetes subjects (Table 2). Pharmacological treatment of the n = 54 patients with type 2 diabetes included insulin, 25%; oral hypoglycemic agents, 66.6%; statins, 58.3%; fibrates, 10.6%; and blood pressure lowering agents, 53.8%. No patients were being treated with thiazolidinedione.
Table 2. Type 2 diabetes cohort: characteristics of the type 2 diabetes cohort
To compare the subcutaneous expression of the gene profile between nondiabetic subjects and patients with type 2 diabetes, we selected 28 nondiabetic and 27 type 2 diabetes subjects. They were matched for BMI, gender, and waist-to-hip ratio (Table 3), thus excluding the potential influence of confounding factors. The pharmacological treatment selected for the 27 patients with type 2 diabetes consisted of insulin, 8%; oral hypoglycemic agents, 77%; statins, 38%; and blood pressure lowering agents, 40%.
Table 3. Type 2 diabetes cohort: groups selected for the gene expression study
The Hospital's Ethics Committee approved the study and informed consent was obtained from all participants.
BMI was calculated as weight (kg) divided by height (m2). Waist circumference was measured midway between the lowest rib margin and the iliac crest. Hip circumference was determined as the widest circumference measured over the greater trochanter. Waist-to-hip ratio was calculated accordingly.
Blood samples were drawn from each subject before breakfast, between 8 and 9 am, after overnight bed rest. Plasma and serum samples were stored at −80 °C until analytical measurements were performed, except for glucose and HbA1c, which were determined immediately after the blood was drawn.
Serum glucose was measured with a glucose oxidase method using a Hitachi auto analyser. Lipid profile (triglycerides, total cholesterol, and high-density lipoprotein cholesterol) was measured using the usual enzymatic methods.
Levels of plasma interleukin-6 (IL-6) were measured by the highly sensitive quantitative sandwich-enzyme-immunoassay technique with the Human IL-6 Quantikine HS ELISA Kit (R&D Systems, Minneapolis, MN). The minimum detectable mean concentration was 0.039 pg/ml. Intra-assay and interassay variation coefficients were <9.8% and <11.2%, respectively. Plasma adiponectin concentrations were measured by radioimmunoassay (Linco Research, St. Charles, MO). The intra-assay and interassay variation coefficients were <5%. Serum insulin was measured by radioimmunoassay (Coat-A-Count Insulin; DPC, Los Angeles, CA). Intra-assay and interassay variation coefficients were <6%. Soluble TNF receptor 1 and 2 were determined by solid phase enzyme immunoassay with amplified reactivity (BioSource Europe, Nivelles, Belgium). The limit of detection was 50 pg/ml for soluble TNF receptor 1 and 0.1 ng/ml for soluble TNF receptor 2, and the intra- and interassay variation coefficients were <7 and <9% respectively.
Adipose tissue samples
Adipose tissue samples were obtained from subcutaneous abdominal depots. Samples were obtained during abdominal elective surgical procedures (gastric bypass operation, cholecystectomy or surgery for abdominal hernia). All patients had fasted overnight, and at the beginning of surgery, 2–4 g of subcutaneous fat tissue was removed from each proband and was immediately frozen in liquid nitrogen and stored at −80 °C until RNA extraction.
Total RNA isolation and reverse transcription
Total RNA was extracted from 400 to 500 mg of frozen subcutaneous adipose tissue using the RNeasy Lipid Tissue Midi Kit (Qiagen Science, Germantown, MD) following the manufacturer's instructions. Total RNA was quantified by absorbance measurement and its purity assessed by the OD260/OD280 ratio.
One microgram of RNA was retrotranscribed to cDNA using the Reverse Transcription System (Promega, Madison, WI) in a final volume of 20 μl, following the manufacturer's instructions.
Real-time quantitative PCR
Primers. The following primers were used: 5′-ggatgaggatggagatgacc-3′ and 5′-gcagaagagccagaggagat-3′ for specific lamin A isoform; 5′-ggatgaggatggagatgacc-3′ and 5′-cacgggggaggctggggagag-3′ for specific lamin C isoform; 5′-ctatggagttcatgcttgtg-3′ and 5′-gtactgacatttattt-3′ for peroxisome proliferator-activated receptor (PPARγ); 5′-aaggtgaagtcggcgcgg-3′ and 5′-atcggggctggcaggg-3′ for SREBP1c; 5′-gagcactgaaagcatgatcc-3′ and 5′-gctggttatctctcagctcca-3′ for tumor necrosis factor α (TNFα); 5′-cggtacatcctcgacgg-3′ and 5′-tgatgattttcaccaggc-3′ for IL-6; and 5′-tctgtgcctgctgctcatag-3′ and 5′-cagatctccttggccacaat-3′ for monocyte chemotactic protein-1 (MCP-1). The housekeeping genes used to normalize gene expression were: β-actin 5′-ggacttcgagcaagagatgg-3′ and 5′-agcactgtgttggcgtacag-3′ and, cyclophilin A (CYPA) 5′-caaatgctggacccaacac-3′ and 5′-gcctccacaatattcatgccttctt-3′. All primers were synthesized by Sigma (Sigma-Genosys, Haverhill, UK).
Real-time PCR. Gene expression analysis was performed on a LightCycler Instrument (Roche Diagnostics, Basel, Switzerland), using the SYBR green fluorescence method. Quantification of 2 μl of the cDNA was carried out in a 20 μl volume of a mixture containing 0.3 μmol/l primers (except for β-actin and TNFα that were 0.5 μmol/l), and 2 μl of LC-FastStart DNA Master SYBR green I (Roche) in a 20 μl volume. The final concentration of MgCl2 was adjusted for each gene.
The purity of each amplified product was confirmed by melting curve analysis and detection of the fluorescent signal was adjusted to avoid primer-dimer detection.
Result analysis. For each sample, the derived gene quantification was calculated from an external standard curve, created with serial dilutions of a cloned PCR fragment from the respective gene, using LightCycler Software v.3.5 (Roche).
Adipose tissue expression levels for each gene were normalized using β-actin. Statistical analysis showed no differences in gene expression for this housekeeping gene among the studied groups.
Sample size was calculated to achieve a difference of 0.3 or greater in logarithm of mRNA mean levels between the studied groups, with a confidence level of 95% and a statistical power of 80%.
Statistical analysis was carried out by using the SPSS/PC+ statistical package (v. 13.5 for Windows; Chicago, IL). Data are expressed as mean ± s.d.
Differences in concentrations and in clinical or laboratory parameters between groups were compared by using either an independent samples t-test or ANOVA, where appropriate. Variables that did not have a Gaussian distribution were logarithmically transformed to perform statistical analysis or were analyzed with nonparametric tests. The differences in sex between the study groups were analyzed by the Pearson/Fisher χ2-test. The relationship between variables was tested using Pearson/Spearman's correlation analysis and stepwise multiple linear regression analysis. Statistical significance occurred if a computed two-tailed probability value was <0.05.
Subcutaneous adipose tissue gene expression in the obesity cohort
The main anthropometrical and analytical characteristics are shown in Table 1. The obese patients were slightly younger than the nonobese subjects and they were predominantly women. All statistical analyses were adjusted for both variables (i.e., age and gender).
The mRNA expression levels of LMNA, adipogenic, and inflammatory genes in subcutaneous adipose tissue are shown in Table 1. In order to assess whether there was any imbalance in the lamin A to C ratio this was also calculated. Severely obese subjects showed a tendency toward higher expression levels of both lamin A and lamin C isoforms; however, only lamin C mRNA levels reached statistical significance. Conversely, lamin A to C ratio was lower in these subjects.
SREBP1c mRNA expression levels were significantly lower in the severely obese group (P < 0.05), and no differences were found in PPARγ mRNA expression according to BMI intervals. The proinflammatory cytokine profile increased in parallel with the BMI (Table 1).
Subcutaneous adipose tissue gene expression in patients with type 2 diabetes
The main anthropometrical and analytical characteristics of this cohort are described in Table 2. Contrary to the subjects without type 2 diabetes, we found no differences in lamin A and C expression according to the BMI intervals in the 54 patients with type 2 diabetes (P = 0.982 and P = 0.176, for lamin A and C).
When comparing gene expression between patients with type 2 diabetes and the nondiabetic population, both lamin A and C isoform mRNA expression were significantly higher in patients with type 2 diabetes, whereas lamin A to C ratio was lower. Significant differences were also found in adipogenic (SREBP1c and PPARγ) and MCP-1 gene expression (Table 4).
Table 4. Type 2 diabetes cohort: gene expression levels (arbitrary units)
Correlation and regression analysis
We performed a linear correlation analysis between the studied genes and the anthropometrical and analytical data including all subjects (n = 106). The independence of the associations was further evaluated by linear multiple regression analysis considering gender and the presence of diabetes as confounding variables.
In bivariate analysis, lamin A mRNA expression levels correlated positively with lamin C (r = 0.645, P < 0.001), IL-6 and MCP-1 mRNA expression (r = 0.337, P = 0.029 and r = 0.480, P < 0.001, respectively), plasma fasting glucose, and triglycerides (r = 0.376, P = 0.001 and r = 0.229, P = 0.023), and negatively with SREBP1c mRNA expression (r = −0.263, P = 0.013) and adiponectin plasma levels (r = −0.389, P = 0.005). In the multiple regression analysis, after adjusting for the confounding variables mentioned above, lamin A mRNA expression was only determined by lamin C expression levels (B = 0.151, P = 0.003, 95% confidence interval CI(B): 0.060/0.243; R = 0.610).
With regard to lamin C mRNA expression levels, a positive correlation was observed with lamin A (r = 0.645, P < 0.001) and with IL-6, MCP-1, TNFα and PPARγ mRNA expression (r = 0.586, P < 0.001; r = 0.506, P < 0.001; r = 0.275, P = 0.006; r = 0.450, P < 0.001, respectively). Likewise, serum insulin levels, plasma fasting glucose, and triglycerides (r = 0.247, P = 0.027; r = 0.416, P < 0.001; and r = 0.209, P = 0.044, respectively) showed a positive correlation. A negative association with SREBP1c mRNA expression (r = −0.361, P = 0.001) was observed. In the multiple regression analysis, lamin A mRNA (B = 2.218, P < 0.001, 95% CI(B): 1.499/2.938); IL-6 mRNA (B = 0.557, P < 0.001, 95% CI(B): 0.392/0.722); and SREBP1c mRNA (B = 0.854, P = 0.001, 95% CI(B): 0.389/1.318) contributed independently to the variability of lamin C mRNA expression (R = 0.939).
As the mRNA expression levels of lamin isoforms seem to be related, we performed a correlation analysis of the ratio between both isoform levels and anthropometrical, clinical, and expression variables. The lamin A to C expression ratio correlated positively with SREBP1c mRNA expression (r = 0.351, P = 0.001) and negatively with fasting glucose (r = −0.327, P = 0.001), cytokine expression: IL-6, MCP-1 and TNFα (r = −0.482, P = 0.001; r = −0.447, P < 0.001; and r = −0.271, P = 0.008, respectively), and with the expression of the adipogenic marker PPARγ (r = −0.449, P < 0.001). In the multiple regression analysis, after adjusting for confounding variables, IL-6 and SREBP1c mRNA expression (B = −0.550, P < 0.001, 95% CI(B): −0.701/–0.400; and B = −0.371, P = 0.001, 95% CI(B): −0.567/–0.175, respectively; R = 0.864) remained associated to LMNA isoform ratio levels.
These results show an upregulation of the lamin A/C mRNA expression in subcutaneous adipose tissue in the context of massive obesity and in type 2 diabetes. When association with lamin A and C expression was analyzed, we found that both isoforms were mutually dependent. Additionally, we also found that adipose lamin C levels were dependent on the expression of IL-6 cytokine and SREBP1c adipogenic marker mRNA expression. In cases where the dependent variable was lamin A to C ratio, regression analysis corroborated these results.
It is known that LMNA expression is increased in highly differentiated cells, whereas their expression is lower in cells with a higher proliferative capacity (27,31,32,33,34). Their expression levels increase during adipocyte (27) and monocyte differentiation (32) and although its role in adipogenesis is as yet unknown, lamin has a function as scaffold for the assembly of other proteins that are involved in adipogenesis (34). Obesity is the result of an increase in adipocyte size and number; adipocytes accumulate triglycerides during their lifetime up to a maximum point, after which proliferation is triggered (35). As severe obesity is the result of a massive expansion of white adipose tissue and requires recruitment of adipocyte precursor cells, severely obese adipose tissue is supposed to carry out more adipogenic processes than in lean subjects. In this sense, the increased levels of lamin A/C mRNA expression observed in severely obese subjects are in accordance with this observation. The observed pattern of expression could reflect an increment in the adipocytic/preadipocytic cell ratio at the expense of an increase in mature adipose cells. Lamin A/C mRNA expression levels increase consistently in subcutaneous preadipocytes from healthy subjects during ex vivo differentiation (27). It has been proposed that the expression of A-type lamins may lock cells into a differentiated state by modifying chromatin organization (36).
An important interaction between lamin A and SREBP1c has recently been described in the context of a disease with an extensive affection of subcutaneous adipose tissue, FPLD. Although the majority of mutations that lead to FPLD did not appear to influence either the expression or the localization of lamin A/C, SREBP1-lamin A binding is reduced (37). In our study, we observed a progressive increase in both lamin A/C mRNA expression with a higher body weight, with significant differences in lamin C in severely obese patients (Table 1). In these severely obese subjects, the subcutaneous expression of SREBP1c is reduced when compared with lean and obese subjects, in accordance with previous studies (38,39,40), suggesting a link between the expression pattern of both genes.
Regarding other adipogenic genes, lamin C was also positively correlated with PPARγ mRNA expression; however, this association disappears after checking for confounding variables. In fact, PPARγ mRNA expression was independent of the BMI, as in other studies, and also in whole subcutaneous adipose tissue, in which PPARγ mRNA levels were found unchanged in both lean and obese subjects, including the severely obese (41,42). On the other hand, in isolated adipocytes from subcutaneous biopsies, variations in PPARγ expression related to BMI have been described (43,44).
Type 2 diabetes patients showed higher expression of lamin A/C mRNA than nondiabetic subjects, independently of the BMI. Although it is worth mentioning that our type 2 diabetes cohort received different treatments, thereafter, it is possible that the increase in lamin A/C expression could be because of the treatment of their type 2 diabetes state rather than the disease itself. We are aware that we are analyzing the whole adipose tissue, in which many different cell types coexist. In this sense, it is known that immune inflammatory cells may also contribute to the expression of lamin A/C mRNA. In fact, we have found lamin A/C mRNA expressed in both stromovascular and adipocyte fractions (data not shown). However, the absence of differences in the inflammatory component of the studied subcutaneous adipose tissue of patients with type 2 diabetes, except for MCP-1, helps us to speculate that the metabolic changes observed in diabetes may induce an increase in lamin A/C mRNA expression. In this sense, the absence of modification in lamin A/C mRNA expression in isolated adipocytes after a hyperglycemic and inflammatory stimulus (M. Miranda, personal observation), in conjunction with increased levels in obese people, suggest that modifications in lamin A/C may be induced by common metabolic pathways involved in both diabetes and obesity before hyperglycemia is established. In fact, nondiabetic carriers of missense mutations in the gene encoding lamin A/C (LMNA) with partial lipodystrophy suffer from hyperinsulinism and show higher free fatty acid and C-reactive protein serum levels (45). It is difficult to extrapolate the metabolic abnormalities with nuclear lamin deregulation, both with a deficit or with an excess of mRNA expression; however, it is known that A-type lamins play a critical role in the regulation of gene expression through the organization of intranuclear chromatin (46). A secondary event in lamin A/C mRNA changes cannot be discounted in the context of our study, as the transversal design does not allow us to confirm whether the observed findings are the consequence of metabolic events rather than the cause. However, standing by our data, the contribution of IL-6 and SREBP1c adipose subcutaneous expression on lamin C levels after multiple regression analysis, independent of BMI or diabetes status, allows for the possibility that lamin C may be modulated by paracrine changes in adipose tissue that occur in both obesity and diabetes.
As the rs4641 polymorphism is located at the site of alternative splicing in exon 10, it is proposed that this splicing variation may change the lamin A to C ratio and thereby prevent the lamin A and C proteins from forming functional polymers, which may affect signaling cascades in adipose and muscle cells (13). This fact, together with the altered lamin A-to-C mRNA expression ratio observed in our study in both obesity and type 2 diabetes, could highlight LMNA as a collaborator in common metabolic phenotypes such as obesity and type 2 diabetes. As Wegner et al. propose, changes in expression or minor genetic variation of LMNA could mirror what happens in more severe LMNA mutations causing monogenic syndromes.
In conclusion, these preliminary results support a possible participation of lamins in the events observed in adipose tissue homeostasis in the setting of obesity and type 2 diabetes. In order to overcome some of the limitations of our study, further work needs to be done on larger cohorts and on treatment-naive type 2 diabetes patients. Studies focusing on mechanisms of action at the molecular level would better help to characterize the role of A-type lamins in both morbidities.
This study was supported by the following grants: REDIMET (RD06/0015/0011), FIS 04-0377, and FIS 05-1994. M.R.C. is supported by a fellowship from the Fondo de Investigación Sanitaria CP06/00119. The CIBER de Diabetes y Enfermedades Metabólicas Asociadas is an initiative of ISCIII.
The authors declared no conflict of interest.