| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
University Hospital for Children and Adolescents (A.K., A.G., R.T., W.K.), Department of Internal Medicine III (M.B.), Institute of Laboratory Medicine, Clinical Chemistry and Molecular Diagnostics (J.K.), University of Leipzig, 04103 Leipzig, Germany
Address all correspondence and requests for reprints to: Dr. Antje Körner, Research Laboratory, University Hospital for Children and Adolescents, University of Leipzig, Liebigstrasse 20a, 04103 Leipzig, Germany. E-mail: antje.koerner{at}medizin.uni-leipzig.de.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Results: Distinct preparations of human visfatin (recombinant, endogenously expressed from human adipocytes, and overexpressed in COS-7 cells) were readily identified by three currently available immunoassays. However, direct comparison of native human serum samples did reveal great discrepancies between these assays and complete lack of correlation. To specify putative molecular isoforms of visfatin, we fractionated iodine-125-labeled recombinant visfatin spiked into human serum and supernatants of visfatin-overexpressing COS-7 cells by size exclusion chromatography. We obtained a distinct peak at approximately 100 kDa that was confirmed by subsequent Western blotting of the fractions and is equivalent to the molecular mass of the dimer. Only one of the immunoassays detected a similar peak in native human size exclusion chromatography serum fractions, whereas the others detected a peak at more than 500 kDa or did not show any distinct peak. We did not observe any differences in visfatin serum levels between lean or obese patients. In addition, there was no correlation between visfatin serum levels with visfatin mRNA expression in sc or visceral fat and with parameters of glucose metabolism.
Conclusion: Differences in the qualitative and quantitative detection of visfatin by immunoassays need to be considered in clinical association studies and may explain the conflicting observations with respect to a putative relation of circulating visfatin to human obesity or insulin resistance.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The effects of visfatin on adipogenesis and glucose metabolism (1) are of particular interest with respect to a putative role in the pathogenesis of obesity and diabetes. Some subsequent clinical studies confirmed the association of visfatin and diabetes (9, 10, 11, 12, 13), whereas others did not find an association (14, 15, 16). Similarly, data regarding the relationship of visfatin with parameters of glucose metabolism and insulin resistance were contradictory (9, 10, 11, 17, 18, 19), and overall, there was no clear effect of visfatin on metabolism. Even more controversial is the discussion on the association of visfatin with obesity with positive (13, 17), negative (14, 20), or lack of association (10, 21), all having been reported. In addition, in contrast to the original publication (1), others found no difference in visfatin mRNA expression between visceral and sc adipose tissue (12, 17).
The molecular and biochemical features of visfatin have not been fully characterized. Currently, it is unknown what molecular form of visfatin occurs in the circulation. Because its enzyme function as Nampt requires dimerization (22), it may circulate as a dimer in human serum. This may have implications with respect to our knowledge on the immunological detection as well as biological actions of visfatin.
We hypothesized that the inconsistencies in clinical studies may potentially be attributed to differences in the specificity of the immunoassays applied. Therefore, we aimed to: 1) assess the molecular characteristics of distinct visfatin protein preparations; 2) evaluate their detection in three currently available immunoassays; and 3) assess visfatin serum levels in lean and obese subjects, and their association with body mass index (BMI), body fat distribution, and glucose metabolism.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Cell culture media were obtained from PAA GmbH (Cölbe, Germany) or Invitrogen GmbH (Karlsruhe, Germany), supplements, antibiotics, and fetal bovine serum from Biochrom AG (Berlin, Germany). Most other reagents were purchased from Sigma-Aldrich (Munich, Germany). Rosiglitazone (BRL 49653) was kindly provided by GlaxoSmithKline (London, UK). Recombinant visfatin produced in Escherichia coli (ALX-201–319) or HEK293 cells (ALX-201–336) was obtained from Axxora (Lörrach, Germany).
Cell culture and overexpression of visfatin in COS-7 cells
COS-7 cells were transiently transfected with a vector containing human visfatin/PBEF (pSR
-PBEF; provided by Dr. A. Fukuhara, Osaka University, Toyonaka, Osaka, Japan) (1). After overnight culture, cells were washed with PBS and incubated in serum-free medium for 48 h.
Human Simpson Golabi Behmel syndrome (SGBS) (23) preadipocytes, kindly provided by Dr. M. Wabitsch (University of Ulm, Ulm, Germany) were grown and differentiated into mature adipocytes as described previously (23, 24). Cell culture supernatants were concentrated 10-fold by lyophilization and reconstituted in PBS.
RT-PCR and quantitative real-time RT-PCR
Total RNA was isolated using RNeasy columns with on-column DNA digestion according to the manufacturer’s protocols (QIAGEN, Hilden, Germany). After RT, visfatin cDNA was amplified using the primers: forward, ATGAATCCTGCGGCAGAAGC; and reverse, CTAATGATGTGCTGCTTCCAGT.
Quantitative real-time PCR (TaqMan) was performed applying kit and protocols by Eurogentec (Cologne, Germany) on the ABI 7500 Sequence Detector (Applied Biosystems Inc., Weiterstadt, Germany) with the following primers: forward, GCAGAAGCCGAGTTCAACATC; reverse, TGCTTGTGTTGGGTGGATATTG; and probe, TGGCCACCGACTCCTACAAGGTTACTCAC for vistatin. Samples were run in triplicates and normalized to a standard curve of serial dilutions of linearized visfatin plasmid (102 to 105 copies). Simultaneous amplification of 18s ribosomal RNA (Applied Biosystems) was used as internal control.
Detection of visfatin by Western blot
Visfatin in supernatants and cell/tissue lysates was detected by Western blot using a monoclonal antibody (OMNI379; Axxora). Cells were lysed in ice-cold 50 mM HEPES lysis buffer (pH 7.5), 150 mM NaCl, 10 mM EDTA, protease-inhibitor cocktail Roche complete (Roche GmbH, Mannheim, Germany), 1% Triton X-100 for 20 min before centrifugation (12,000 x g, 10 min, 4C). After electrophoresis and blotting, membranes were incubated in 5% dry milk in Tris-buffered saline containing 0.1% Tween 20. After incubation with anti-visfatin primary antibody overnight (1:2500), membranes were incubated with horseradish peroxidase-goat-antirabbit IgG (Dako GmbH, Hamburg, Germany) and visualized using the ECL detection system (SuperSignal West Pico, Pierce Biotechnology, Inc., Rockford, IL). For negative control, anti-visfatin antibody was substituted by equimolar normal mouse IgG2a (SCBT, Heidelberg, Germany). Recombinant human IgG (75 µg) was not detected by anti-visfatin antibody. Cell culture supernatants were concentrated 10-fold by lyophilization and reconstituted in PBS.
Detection of visfatin by immunoassays
Three commercially available immunoassays comprised of an enzyme immunoassay (EIA) (Phoenix Europe GmbH, Karlsruhe, Germany), RIA (Phoenix), and ELISA (AdipoGen Inc., Seoul, Korea) were performed according to the manufacturer’s instructions.
The EIA, a competitive EIA, used a polyclonal rabbit antibody directed against a C-terminal peptide that also served as standard protein of the assay. The detection range was 0.1–1000 ng/ml with an interassay variation of less than 14% and intraassay variation less than 5% according to the manufacturer.
The RIA is a competitive RIA using a purified polyclonal antibody. Iodine-125 (125I)-labeled recombinant visfatin was used as standard protein. The detection range was 15.6–2000 ng/ml.
The ELISA, a sandwich ELISA, used a monoclonal immobilized antibody and a polyclonal rabbit antihuman visfatin capture antibody. Recombinant human visfatin expressed in HEK293 cells served as standard protein. The detection range was 0.03–16 ng/ml. Interassay and intraassay variation was given with 6.31–9.53 and 3.46–5.53%, respectively. The assay had 100% cross-reactivity with rat visfatin and 20% cross-reactivity with mouse visfatin. Recovery rates from human serum were given between 89.1 and 109.7%, and serial dilution of human serum samples were 83.8–104.4% of expected levels, according to the manufacturer’s information. Cross-reactivity with other adipocytokines was 0% for all three assays, according to the manufacturers.
Size exclusion chromatography (SEC)
SEC experiments were performed using a Superdex 200 column (16/60) (General Electrics Healthcare, München, Germany) in PBS buffer. A sample of 0.9 ml serum was subjected to the column and eluted with a flow rate of 1 ml/min at 4 C. The distribution of molecular masses within the eluates of the column was adjusted by measuring the absorbance of commercially available calibration proteins (Pharmacia-Biotech, Freiburg, Germany).
The 125I-labeled visfatin tracer of the RIA (30,000 cpm, 0.12 µCi/ml) was used for spiking experiments with 450 µl human serum and the subsequent separation by SEC. In a second run, an excess of 600 ng recombinant visfatin was added to displace the labeled visfatin from potential binding proteins.
Subjects and collection of serum samples
A total of 30 serum samples was selected from subjects of a larger cohort described previously (17) to provide a wide range of BMI and insulin sensitivity determined by euglycemic-hyperinsulinemic clamp (Table 1
). Subjects were newly diagnosed as having normal glucose tolerance or type 2 diabetes on the basis of fasting and 2-h oral glucose tolerance test plasma glucose concentrations, and were classified as sc (n = 10) or visceral (n = 10) obese on the basis of fat distribution on a computed tomography scan. Subjects had never been treated with oral antidiabetic medications or insulin. Patients with severe conditions, including generalized inflammation or end-stage malignant diseases, were excluded from the study. Samples of visceral and sc adipose tissue were immediately frozen in liquid nitrogen. Visfatin mRNA expression was assessed as described previously (17). Blood samples were collected after overnight fasting. Measurements were performed after a maximal storage time of 1 wk on samples that were thawed and analyzed simultaneously. The study was approved by the ethics committee of the University of Leipzig. All subjects gave written informed consent before participating in the study.
|
For all cell biology experiments, at least three independent cell culture experiments were performed. Data are presented as mean ± SEM. Data were log transformed if they did not adhere to normal distribution. Differences between means were analyzed by the Student’s t test. Correlation analyses were performed by Pearson’s correlation analysis.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We detected visfatin mRNA expression in paired samples of visceral and sc human adipose tissue, as well as in preadipocytes and adipocytes of the SGBS cell line (Fig. 1A). Quantification of visfatin mRNA expression did not reveal a distinct pattern during adipocyte differentiation, even though expression appeared lower in preadipocytes (Fig. 1B).
|
). Corresponding to the mRNA expression, we detected visfatin protein in white adipose tissue and adipocyte cell lysates during all stages of differentiation, whereas detection of visfatin protein in the supernatant was predominant in differentiated adipocytes (Fig. 1D), as described previously (22). Overexpression of human full-length visfatin cDNA in COS-7 cells resulted in an increased mRNA (Fig. 1A) and protein expression in cell lysates and supernatants (Fig. 1E). Thus, we characterized three distinct sources of visfatin, namely commercially available prokaryotic and eukaryotic visfatin, endogenously expressed visfatin from human adipocytes, and visfatin overexpressed and secreted from transfected COS-7 cells.
Detection of visfatin by immunoassays
Applying the visfatin protein preparations from these three distinct sources, we aimed to verify visfatin detection by the three immunoassays that are currently available: a competitive EIA with an antibody directed against a C-terminal peptide, a competitive RIA, and a sandwich ELISA. In the following, these assays are referred to as EIA, RIA, and ELISA for reasons of simplicity.
All three assays detected serial dilutions of prokaryotic and eukaryotic recombinant protein, although the magnitude of measured concentrations was different (Fig. 2A).
|
), whereas the ELISA did not identify visfatin in control media in correspondence with the Western blot results. In the supernatants of COS-7 cells overexpressing the visfatin transcript for 48 h, we obtained a strong signal with all assays applied (Fig. 2C). Therefore, all three assays detected defined preparations of eukaryotically expressed visfatin protein.
However, when we subsequently compared visfatin concentrations in native human serum samples, there was no correlation between visfatin serum levels obtained by the EIA and RIA or between the EIA and ELISA (Fig. 3, A and B). Bland-Altman analyses visualized the great discrepancies with 2-fold differences at lower concentrations that increased up to 5- to 9-fold differences in higher concentration ranges when comparing the RIA and the EIA (Fig. 3C). The discrepancies were even higher between the ELISA and EIA with differences over several magnitudes at an average concentration of 20 ng/ml (Fig. 3D).
|
To specify putative molecular isoforms of visfatin, we fractionated supernatants of visfatin-overexpressing COS-7 cells by SEC obtaining 30 serum fractions with defined molecular masses ranging from approximately 20 to more than 800 kDa.
In the fractionated supernatants of visfatin-overexpressing COS-7 cells, we obtained a distinct peak with a maximum at approximately 100 kDa that was confirmed by simultaneous Western blotting (Fig. 4A). To confirm the molecular mass and assess putative binding of visfatin to binding proteins in human serum, we similarly subjected 125I-labeled recombinant visfatin spiked into human serum to SEC with subsequent displacement by unlabeled visfatin. Again, a distinct peak of activity was detected at approximately 100 kDa, with a smaller shoulder at approximately 50 kDa, hence, corresponding to the calculated molecular mass of the dimer and monomer, respectively (Fig. 4B). From these data it is conceivable that visfatin in human serum occurs as a dimer.
|
The incongruent results in regard to visfatin detection in human serum samples obtained with different immunoassays (Fig. 3) led us to suspect that there might be distinct compounds detected by the different immunoassays.
None of the assays showed cross-reactivity with recombinant human insulin. Recombinant human IgG (10 mg/ml) subjected to the immunoassays gave low signals in all immunoassays that were, however, lower than the minimal visfatin serum levels: IgG (visfatin serum levels range); EIA, 19.0 (35.4-57.2); RIA, 24.0 (98.0-142); ELISA, 0.284 (0.915-1.614). The recovery rates for 10 ng recombinant eukaryotic visfatin spiked into human serum were 156% for the EIA and 118% for the ELISA, respectively.
We next aimed to confirm the detection of the visfatin dimer in SEC-fractionated human serum samples by the immunoassays. Applying the ELISA, we obtained a single distinct peak with a maximum at approximately 100 kDa (Fig. 4C). With the RIA, all signals were near the detection limit, and no distinct peak could be specified (Fig. 4D). Applying the EIA, we detected a single distinct peak of high molecular mass more than 500 kDa, but no peak at 100 or 50 kDa (Fig. 4E). In the corresponding Western blot analysis of alternate serum fractions, visfatin was detected in serum fractions with molecular masses between 100 and 150 kDa (Fig. 4F), similar to the ELISA signals, whereas there was no detection corresponding to the high molecular mass peak seen in the EIA. From these data it appears that visfatin in human serum is most specifically detected by the ELISA, whereas the EIA appears to detect some high molecular mass protein, and the RIA is limited by the low sensitivity and/or discrimination.
Association of serum visfatin in lean and obese subjects and association with white adipose tissue mRNA expression
Considering that almost all clinical studies investigating the association of serum or plasma visfatin levels with obesity and type 2 diabetes were performed with the EIA that, according to our results, appears to be limited in specificity, we evaluated visfatin serum levels in 10 sc obese, 10 visceral obese patients, and 10 lean controls applying the ELISA. Parameters of obesity, glucose metabolism, and adipocytokines were significantly different in the obese groups compared with the lean controls, whereas there were no differences in age (Table 1).
Visfatin serum levels were not significantly different between female and male subjects (2.06 ± 1.89 ng/ml, n = 19 vs. 1.79 ± 1.71 ng/ml, n = 11). Although there appeared to be a tendency for increased visfatin serum levels in visceral obese patients, this was not statistically significant (Fig. 5A). In addition, there was no correlation with BMI (Fig. 5B) or anthropometric parameters [waist circumference, hip circumference, waist-to-hip ratio (WHR), body fat], nor with glucose metabolism (fasting blood glucose, fasting plasma insulin, glucose infusion rate, 120-min blood glucose in oral glucose tolerance test), with plasma lipids nor was there any correlation with visfatin mRNA expression in sc or visceral fat depots.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Our data show that visfatin is released as a dimer from PBEF-transfected COS-7 cells as a model of eukaryotic visfatin overexpression. To identify putative visfatin molecular isoforms or aggregation with binding proteins in human serum, we applied a similar approach by subjecting iodine-labeled visfatin that was spiked into human serum to SEC. Again, the strongest signal was obtained at approximately 100 kDa, with a smaller shoulder at 50 kDa, equivalent to the molecular mass of the dimer and monomer, respectively. These data strongly suggest that the majority of visfatin protein circulates as a dimer in human serum, as can be expected from their predicted molecular structure (25). The dimerization of visfatin may have functional implications, particularly in regard to its NAD+ biosynthetic enzyme function that has been shown to depend on dimerization and to be active also in extracellular space (22). The absence of additional high molecular mass peaks in the SEC experiments and corresponding Western blots excludes multimerization or the presence of binding proteins to a great extent.
The molecular characteristics of visfatin need to be considered for the applicability of immunoassays and interpretation of their results. In this study we provide evidence of considerable qualitative and quantitative discrepancies in the detection of visfatin by three commercial immunoassays.
We obtained distinct preparations of recombinant, endogenously expressed visfatin from SGBS adipocytes, and overexpressed human visfatin from PBEF-transfected COS-7 cells that we subsequently subjected to immunoassays for validating their accuracy. All assays detected the recombinant compounds as well as visfatin protein in the supernatants of visfatin-overexpressing COS-7 cells, therefore, verifying immunoreactivity against visfatin.
Unexpectedly, the direct comparison of native human serum samples with the three available immunoassays did reveal a complete lack of correlation between visfatin concentrations determined by the EIA and RIA, as well as between the EIA and ELISA. This indicated that the immunoreactivity measured could be directed against interfering as well as cross-reacting compounds or potential isoforms of visfatin. Therefore, we examined what molecular compounds are detected by the immunoassays in SEC fractions of human serum.
The EIA, currently the most widely applied immunoassay, detected visfatin with one distinct peak in the high molecular mass fractions that did, however, not correspond to the detection in Western blots. Possible explanations for a high molecular mass complex may include multimerization of visfatin or binding to other (macro)molecules, as has been shown for other hormones (26, 27). This is unlikely to be the cause, since visfatin has occurred as a dimer in our experiments and previous studies (25, 28). In addition, Western blotting of corresponding serum fractions, which involves a denaturing SDS-PAGE step, did not detect visfatin immunoreactivity at 50 kDa or at higher molecular mass in the respective high molecular mass serum fractions, which strongly argues against these explanations. Therefore, although the EIA did detect the recombinant proteins, interference by proteins other than the visfatin monomer or dimer in human serum is probable. In this regard it should be considered that the primary antibody of this EIA is directed against a C-terminal peptide (also used for the standard curve) and, thus, bears a higher risk of nonspecific binding.
The RIA and ELISA might constitute alternative tools for detecting visfatin. Concerning the RIA, there were considerable limitations in the sensitivity of visfatin detection in human serum samples that may also account for the absence of a distinct peak in the SEC fractions. Of the three immunoassays, the ELISA was the only assay that detected visfatin in SEC fractions of human serum at the expected molecular mass of approximately 100 kDa. However, this assay did show some limitations due to the narrow detection range (0.25–16 ng/ml).
Thus, our results identified the ELISA as the most specific and, therefore, applicable assay for visfatin detection in human serum, whereas in our hands, the EIA and RIA did not reliably quantify visfatin in human serum samples. These observations have implications on the interpretation of the controversial studies that mostly used the EIA. We assessed the association of serum visfatin, measured with the ELISA, with obesity and glucose parameters but did not find any differences between lean and obese subjects, either sc or visceral obese. There was also no correlation between visfatin serum levels and glucose tolerance or any other metabolic parameter. One limitation of our study may be the small samples size. Nevertheless, we did identify expected and significant differences not only in BMI, but also in leptin, adiponectin, and IL-6 levels between lean controls and the obese groups. In addition, the data sample was sufficient to confirm significant correlations of leptin and adiponectin with BMI, WHR, insulin, and blood glucose infusion rate during clamp studies. Overall, our study does not support the hypothesis that visfatin is increased in visceral obesity in parallel with insulin resistance (1), although some studies did show associations with diabetes (9, 10, 13, 18, 19). However, our results are in accordance with other reports that failed to confirm the association of visfatin with obesity (10, 14, 20, 21) and with parameters of glucose metabolism or insulin resistance (10, 11, 17, 18, 19). The methodological shortcomings in the immunoassays may explain, at least in part, the inconsistencies and, if at all, subtle associations of circulating visfatin levels with obesity or insulin resistance. Other explanations for the variability include different sample additives, storing conditions and times, and other preanalytical conditions (29). Overall, visfatin serum levels do not appear to constitute a major marker of visceral obesity and impaired glucose tolerance.
Eventually, a provocative explanation would be to question whether visfatin is a freely circulating adipocytokine (30). Visfatin is clearly expressed by adipocytes on the mRNA and protein level as shown in this and other studies, but the proposed predominance in visceral fat was not confirmed (12, 17). In addition, expression is not restricted to adipocytes but is high in human liver, muscle, and macrophages (2, 31). Although visfatin protein lacks a typical signal sequence (2) and was predominantly found intracellularly (8), we did detect visfatin protein in the supernatant of differentiated adipocytes in parallel with adiponectin (22).
Nevertheless, the lack of correlation of visfatin serum levels with obesity or metabolic parameters does not exclude a role of visfatin in glucose metabolism. In its function as an NAD+-synthetic enzyme, visfatin/Nampt is suggested to be a crucial component of cellular intermediary metabolism (32). Although, in our hands, visfatin/PBEF/Nampt did not exert insulin-like actions or insulin-signaling capacities, it is involved in glucose metabolism at the β-cell level by regulating insulin secretion via the NAD+ pathway (22).
In conclusion, our results reveal major inconsistencies and limitations in the detection of serum visfatin levels by different immunoassays. This needs to be considered when interpreting data from clinical studies, and it may in part explain the controversial observations on the relation of circulating visfatin to human obesity or insulin resistance. In our cohort there was no association of visfatin with obesity or metabolic parameters. The development of immunoassays that reliably detect visfatin as well as functional studies, including the enzymatic activity, will be desirable to enhance our understanding of visfatin biology and its physiological relevance.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Disclosure Statement: The authors have nothing to declare.
First Published Online September 18, 2007
1 A.K. and A.G. contributed equally to this work. 
Abbreviations: BMI, Body mass index; EIA, enzyme immunoassay; 125I, iodine-125; NAD, nicotinamide adenine dinucleotide; Nampt, nicotinamide 5-phosphoribosyl-1-pyrophosphate transferase; PBEF, pre-B cell enhancing factor; SEC, size exclusion chromatography; SGBS, Simpson Golabi Behmel syndrome; WHR, waist-to-hip ratio.
Received June 12, 2007.
Accepted September 7, 2007.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
2-macroglobulin as a growth hormone-binding protein in human blood. J Clin Endocrinol Metab 80:585–590[Abstract]This article has been cited by other articles:
 |
|
 |
|
  K.-D. Nusken, M. Petrasch, M. Rauh, W. Stohr, E. Nusken, H. Schneider, and J. Dotsch Active visfatin is elevated in serum of maintenance haemodialysis patients and correlates inversely with circulating HDL cholesterol Nephrol. Dial. Transplant., September 1, 2009; 24(9): 2832 - 2838. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Hofso, T. Ueland, H. Hager, T. Jenssen, J. Bollerslev, K. Godang, P. Aukrust, J. Roislien, and J. Hjelmesaeth Inflammatory mediators in morbidly obese subjects: associations with glucose abnormalities and changes after oral glucose Eur. J. Endocrinol., September 1, 2009; 161(3): 451 - 458. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K.-D. Nusken, J. Dotsch, M. Rauh, W. Rascher, and H. Schneider Uteroplacental Insufficiency after Bilateral Uterine Artery Ligation in the Rat: Impact on Postnatal Glucose and Lipid Metabolism and Evidence for Metabolic Programming of the Offspring by Sham Operation Endocrinology, March 1, 2008; 149(3): 1056 - 1063. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
