Introduction

White adipose tissue has, in the last two decades, emerged as an active organ with a pathophysiologic role in insulin resistance, dyslipidemia, and atherosclerosis (1). It is, therefore, not surprising that studies of adipocyte function have been the focus of an ever-increasing number of researchers. Adipocytes have the capacity to store energy in the form of triglycerides (TGs),¹ which are hydrolyzed through lipolysis into free fatty acids and glycerol. The rate-limiting enzyme of this process is hormone sensitive lipase (HSL). There is a continuous hydrolysis of TGs in fat cells termed basal lipolysis; however, this process can be greatly enhanced by catecholamines. Moreover, fat cells secrete a number of soluble factors into the circulation, often referred to as adipokines, of which the most well-known are leptin and adiponectin (2). Through the release of these and other factors, adipose tissue can, in effect, be defined as the largest endocrine organ in mammals.

We and others have used human adipocytes for metabolic research, but the lack of an immortalized human adipocyte cell line is a substantial drawback in the field. In the absence of a human fat cell line, many researchers have, therefore, resorted to studies on immortalized murine cell lines, e.g., 3T3-L1 and 3T3-F442. However, there are important differences between human and murine fat cells (3). For example, human adipocytes express significant levels of the adrenergic receptor 2A (2A-AR), which, in contrast to -adrenoceptors (-ARs), activates anti-lipolytic pathways through inhibition of adenylate cyclase (4). Moreover, human fat cells display a unique prolipolytic response to atrial natriuretic peptide (ANP), an effect that is mediated through guanylate cyclase receptors-A (GC-A) (5). Because of the lack of immortalized human fat cell lines, alternative in vitro systems have been developed. The most common one is to isolate fibroblast-like adipocyte precursors cells (preadipocytes) or mesenchymal stem cells (MSCs) (6, 7) from the stroma-vascular portion of human white adipose tissue and differentiate them in vitro into adipocytes. These methods are hampered by different drawbacks. Availability of both preadipocytes and MSCs is dependent on surgical procedures, and the survival of preadipocytes is limited to a few weeks and they are difficult to cryopreserve.

During the course of this study, several groups have presented results on the derivation of adipocytes from embryoid bodies of human embryonic stem cells (hESCs) (8, 9, 10). In these reports, the authors assessed differentiation by morphologic determination and gene expression. In addition, in the report by Xiong et al. (10), a rather low level of leptin secretion was shown. However, there was no functional assessment of these cells regarding lipid turnover, the most characteristic feature of mature adipocytes. We have derived 25 hESC lines in our laboratory, the Fertility Unit of the Karolinska University Hospital Huddinge (11, 12). These human cell lines have been cultured for several years and extensively characterized (11). In brief, these cells express all markers of pluripotent hESCs. including Oct-4, SSEA-4, TRA-1–60, TRA-1–81, GCTM-2, and alkaline phosphatase, and they display pluripotency in vitro after being grown into embryoid bodies. In addition, they form teratomas when injected into immuno-incompetent mice testes.

In this study, we differentiated hESCs into adipocytes using a protocol similar to those described above (8, 9, 10). Because our primary aim was to use hESC cultures in their fully differentiated state, gene expression for a set of genes specific for terminal adipogenic differentiation was analyzed. Moreover, in contrast to previous studies, we performed a functional assessment on lipolytic capacity.

Research Methods and Procedures

hESC Cultures

All of our hESC lines were derived from supernumerary blastocyst stage embryos, which were donated by couples undergoing in vitro fertilization treatment. The couples all gave their informed consent. We had approval from the ethics committee of Karolinska Institutet for derivation and differentiation of hESCs. In this study, we used HS306, a cell line grown since February 2004 with the karyotype 46XX, and HS346, a cell line grown since April 2004 with the karyotype XX. For derivation and non-differentiated growth of the hESCs, we used postnatal human skin fibroblasts as feeder cells. The culture medium consisted of KnockOut-Dulbecco's modified Eagle's medium (DMEM) (Gibco; Invitrogen, Carlsbad, CA) supplemented with 15% knock-out serum replacement (Gibco), 2 mM L-glutamine, 0.1 mM minimum essential medium non-essential amino acids, 50 U/mL penicillin, 50 g/mL streptomycin, and 8 ng/mL basic fibroblast growth factor (Sigma, St. Louis, MO). The cells were cultured at 37°C in 5% CO₂. For passage at 5- to 7-day intervals, colonies were mechanically dissected into small pieces and replated on a -irradiated (35 Gyr) human foreskin fibroblast feeder layer, and the medium was changed every day as described (11, 12). For differentiation, the cells were removed from the feeder layer, and the basic fibroblast growth factor was not added to the medium any more. The cells were allowed to attach to the cell plate bottom and to grow as embryonic bodies (EBs).

Adipogenic Differentiation of EBs

Adipogenic conversion was induced with specific agents, and a schematic representation of the differentiation protocol is given in Figure 1. Each EB was grown for 7 days in growth medium. Growth medium consisted of DMEM/Ham's F12 media (F12) Glutamax I, 1 g/liter glucose, 100 mg/liter penicillin/streptomycin, and 10% fetal calf serum (FCS). Differentiation medium termed standardized adipogenic medium (SAM) was added at Day 7, and cells were allowed to differentiate and were thereafter maintained for several weeks in culture. SAM consisted of DMEM/F12 Glutamax I, 4.5 g/liter glucose, 100 mg/liter penicillin streptomycin, 0.5 mM isobutyl methylxanthine, 1 M dexamethasone, 10 g/mL insulin, 10% FCS, and 1 M rosiglitazone.

Figure 1.

Schematic diagram of the differentiation protocol. Culture days are numbered from the first appearance of EBs (Day 0).

Full figure and legend (40K)

Isobutyl methylxanthine, rosiglitazone, and dexamethasone were added to the cells only for 2 days. We determined the effect of a serum-free medium where FCS was exchanged for serum replacement (SR; Invitrogen, Carlsbad, CA), whereas all other components remained identical to that of SAM. SR from several different batches was used. We did not observe any significant adipogenic difference between batches. In murine ESCs, a short-term treatment with retinoic acid (RA) before the differentiation process significantly improved adipogenic conversion (13). In our cultures, a 3-day exposure with two different concentrations of RA (10^-7 and 10^-8 M) had a negative effect, because approximately one half of the EBs did not survive (Table 1). In the surviving EBs, no major effect of RA was observed on gene expression or adipokine release (data not shown). Oil Red O stains were performed as described previously (14).

Table 1. - Survival of EBs in various differentiation conditions, starting with eight ESCs.

Full table

Real-time polymerase chain reaction

Total RNA was extracted from each EB using the Nucleospin RNA II Kit (Macherey-Nagel, Düren, Germany), and determination of RNA purity was performed using an Agilent 2100 Bioanalyzer (Agilent Technologies, Kista, Sweden) as previously described (15). RNA was reverse transcribed using the iScript cDNA synthesis kit (Bio-Rad, Hercules, CA). Real-time polymerase chain reaction was carried out on an iCycler iQ Real-time PCR Detection System (Bio-Rad). Primers used were as follows: peroxisome proliferator-activated receptor (PPAR)-2B, sense ACAAGGCCATTTTCTCAAACG; antisense CGGAGAGATCCACGGAGC; 2-AR, sense GCTATGTCAATTCTGGTTTCAATCCC; antisense GTGTTGCCGTTGCTGGAGTAG; 2-AR, sense GTTCGTGGTGTGCTGGTTC; antisense: TGGTTGAAGATGGTGTAGATGAC; HSL, sense CTCAGTGTGCTCTCCAAGTG; antisense CACCCAGGCGGAAGTCTC; 18S, sense TGACTCAACACGGGAAACC; antisense TCGCTCCACCAACTAAGAAC. GC-A mRNA was detected using a Taqman gene expression assay (HS 00181445_M1; Applied Biosystems, Foster City, CA). Samples were normalized using 18S rRNA, and relative quantification of mRNA was calculated using the "Comparative Ct method" as described in User Bulletin 2 from Applied Biosystems.

Adipokine Secretion

In brief, medium was collected from four EBs, pooled, lyophilized, and dissolved in 0.3 mL of distilled water. An aliquot of 100 L was analyzed according to the protocol. Results were corrected for the concentration factor. Leptin was measured using the Quantikine Human Leptin Immunoassay kit (R&D Systems, Minneapolis, MN). Adiponectin was measured using a commercially available radioimmunoassay kit from LINCO Research (St. Charles, MO).

Assessment of Glycerol Release

Glycerol release (basal lipolysis) from the EBs was measured during the differentiation process at each time-point when the medium was changed. Glycerol was measured using a luminometric assay, as described (16). Glycerol release was expressed as nmoles per EB per 24 hours. We also studied stimulated lipolysis on fully differentiated hESC-derived adipocytes. In these experiments, the medium was changed to DMEM/F12 with 0.25% bovine serum albumin, and the EBs were further incubated for 4 hours at 37 °C. An aliquot of the medium was removed for the measurement of glycerol (basal lipolysis). Medium containing either 10^-7 M isoprenaline, 10^-5 M noradrenaline (NA), 10^-5 M NA + 10^-4 M yohimbine (NA+Y), or 10^-7 M ANP was added. The EBs were further incubated for 4 hours, after which aliquots were obtained for the measurement of glycerol (stimulated lipolysis).

Human MSC Culture

Human MSCs were isolated and cultured, and lipolysis assays were performed as described in detail previously (14). In one experiment, we determined stimulated lipolysis using the same reagents and concentrations as above.

Results

Comparison between Serum and SR on Adipogenic Differentiation of HS306

HS306 cells were allowed to form independent cell masses, which are EBs that are capable of differentiating into different cell lineages (11). Adipogenic conversion of murine ESC-derived EBs has previously been described (13). We used a growth and differentiation protocol using SAM supplemented with either fetal calf serum (serum) or SR. We also tested two differentiated EBs in SAM that contained neither FCS nor SR. Under these conditions, none of the EBs survived (data not shown). With SAM supplemented with either serum or SR, EBs differentiated into cells displaying morphologic characteristics of mature adipocytes that could be maintained for 6 weeks in culture. However, there was a clear difference in the appearance of the cells in the two media, because cells grown in the presence of SR displayed larger intracellular lipid droplets (Figure 2). Moreover, cells grown in SR had significantly higher PPAR2B mRNA expression (Figure 3A) and basal glycerol release (Figure 3C) compared with cells cultured in the presence of serum. We also measured the release of the adipokines leptin and adiponectin. The amounts released from any individual EB were too low to detect with our ELISA-based method. We, therefore, pooled conditioned media from four independently grown EBs and assayed protein levels as described in Materials and Methods. This was performed twice in separate experiments. In agreement with the results on mRNA and glycerol release, the levels of both leptin and adiponectin tended to be higher in the SR compared with serum cultured cells (Figure 3B). Taken together, these results suggest that SAM + SR is a better differentiation cocktail than SAM supplemented with serum. All further experiments were, therefore, performed in the presence of SR.

Figure 2.

Influence of media on morphologic differentiation of HS306-derived EBs. Representative photomicrographs of Oil red O-stained EBs grown in (A) serum or (B) SR. Note the enhanced adipogenic differentiation of EBs differentiated in SR. Photos were taken when cells were fully differentiated at Day 37.

Full figure and legend (480K)

Figure 3.

Improved differentiation in SR compared with serum cultured EBs. (A) Expression of the adipocyte-specific PPAR2B mRNA was significantly higher in EBs cultured in the presence of SR than in cells incubated with serum. Data were based on mRNA from four EBs. (B) Adipokine release was studied as indicated. Media from four EBs were pooled together to obtain detectable levels. Each bar represents the mean from two independent experiments containing four individual EBs each. (C) Basal lipolysis, assessed as glycerol release into the medium, was significantly higher in SR-cultured cells. Data were based on medium from four EBs. All experiments were performed when cells were fully differentiated at Day 37 (* p < 0.05).

Full figure and legend (74K)

Lipolysis in hESC-derived Adipocytes

The spontaneous release of glycerol is a hallmark of mammal adipocytes and, in contrast to mRNA expression, a sensitive functional marker of adipogenic conversion. We measured basal glycerol release and found that it increased during differentiation and plateaued at 20 days after the start of differentiation (Figure 4A). To determine whether hESC-derived adipocytes express genes involved in lipolysis, we measured a set of mRNA by quantitative reverse transcriptase-polymerase chain reaction. HSL, adrenergic receptors (2-AR, 2-AR), and GC-A were all detectable in hESCs (Figure 4B). This shows that hESC-derived adipocytes express important genes involved in lipolysis. To see whether the differentiated cells displayed functional lipolysis comparable with human adipocytes, the cells were stimulated with the natural catecholamine NA. In contrast to the well-established effect in primary human adipocytes, there was no detectable increase in glycerol release after incubation with NA in hESC-derived adipocytes (Figure 4C). The presence of the anti-lipolytic 2-AR, which is also activated by NA, could putatively explain this finding. However, co-incubation with 10^-4 M yohimbine (an 2-AR inhibitor) did not improve NA-induced lipolysis (Figure 4C). Furthermore, incubation with the -AR selective agonist isoprenaline induced only a minor (10%), albeit significant, increase in glycerol release (Figure 4C). This effect was considerably lower than the 4- to 8-fold increase commonly observed in mature adipocytes or preadipocytes (17). In contrast, somewhat unexpectedly, ANP induced a marked 3-fold increase in glycerol release, comparable with levels found in primary adipocytes (5).

Figure 4.

Basal and stimulated lipolysis of HS306 EBs. EBs were incubated in adipogenic medium for a period of up to 6 weeks. Basal lipolytic activity was determined by assessing glycerol release into the media/24 hours. (A) EBs displayed significantly improved basal lipolysis 13 days after initiation of differentiation that was maintained during the entire culture period. (B) Fully differentiated (Day 37) EBs were lyzed, and mRNA expression of the indicated genes was determined by quantitative real time-polymerase chain reaction. (C) Differentiated EBs were incubated with NA, NA + Yohimbine (NA+Y), isoprenaline, or ANP for 4 hours. Lipolysis is expressed as percentage of basal glycerol release (n = 4, * p < 0.05).

Full figure and legend (84K)

Comparison of HS306-derived Adipocytes with Other Cell Types

To confirm that the observed lipolytic responses were specific, we used the same set and concentrations of prolipolytic compounds in adipocytes derived from hMSCs. This experiment was performed in parallel with the assays on HS306-derived EBs. In hMSC-derived adipocytes, NA displayed a 2-fold increase in glycerol release, which was significantly enhanced in the presence of yohimbine (NA+Y; Figure 5 A). The response to isoprenaline was similar to NA+Y and in the same range as that observed with ANP (Figure 5A). Finally, to assess whether the results obtained in HS306 were specific to this cell line, we performed a comparison with adipocytes derived from the hESC-line HS346. Similar to findings in HS306, these cells displayed basal glycerol release (Figure 5B). Moreover, HS346 cells displayed no significant response to NA or NA+Y but a significant increase in lipolysis on stimulation with ANP.

Figure 5.

Comparisons of lipolysis in hMSC-derived adipocytes and HS346. (A) Glycerol release was determined in hMSC-derived adipocytes using the same reagents as described in Figure 4C . In contrast to findings in HS306, these cells displayed a clear response to NA, NA+Y, and isoprenaline. The effect of ANP was in a similar range. (B) Basal glycerol release was determined in HS346 during the entire differentiation process. (C) The lipolytic response in HS346 was determined with NA, NA+Y, and ANP. These results were very similar to those obtained in HS306. Data were based on medium from four EBs. Experiments were performed when cells were fully differentiated at Day 35.

Full figure and legend (123K)

Discussion

The putative future clinical application of hESCs in stem cell therapy is of great interest but implies that the cells can be directed under controlled and reproducible conditions into different lineages and cell types. Thus far, many researchers have described adipogenic conversion of hESCs. These studies have focused primarily on gene expression analysis and TG accumulation. However, a functional adipocyte should also display a basal capacity to hydrolyze TGs, as well as the capacity to hydrolyze TGs in response to hormonal stimulation. In this study, we studied hESC-derived adipocytes, and in contrast to previous studies, we present a functional assessment during and after adipogenic conversion of these cells. Although the differentiated hESC displayed morphologic and functional similarities to differentiated human adipocytes, we also observed key differences between these cells and mature human adipocytes.

In this study, adipocytes differentiated from hESCs showed important features of mature adipocytes, including mRNA expression for a number of lipolytic-specific genes (HSL, GC-A, and ARs) and secretion of adipokines (leptin and adiponectin). These cells also displayed a basal release of glycerol, which was detected early in differentiation and was maintained throughout the differentiation process. However, although the cells expressed significant mRNA levels of adrenergic receptors and HSL, there was a significantly blunted response to catecholamines. There was no effect of NA, which is in sharp contrast to findings in freshly isolated human adipocytes or differentiated preadipocytes. This was clearly not due to a strong activation of the anti-lipolytic 2-AR, because yohimbine did not affect NA-induced lipolysis, and the -AR selective agonist isoprenaline had only a small (10%) lipolytic effect. At present, we have no clear explanation for this low catecholamine response. It is not caused by insufficient concentrations or degradation of the agonists because the same concentration of reagents resulted in efficient lipolytic activation in adipocytes derived from human mesenchymal stem cells (hMSCs). The differentiation medium with both serum and SR contained insulin at supraphysiologic concentrations, which may inhibit catecholamine action by increasing phosphodiesterase 3B activity and, hence, cyclic adenosine monophosphate hydrolysis. However, all lipolytic assays were performed after washings and in the absence of insulin. Moreover, in human preadipocytes or hMSC-derived adipocytes, we observed a significant catecholamine response, even in the presence of high insulin concentrations. In contrast to the findings with catecholamines, the lipolytic response to ANP in hESC-derived adipocytes was in a similar range to that observed in primary human adipocytes (5). ANP-induced lipolysis is dependent on activation of the guanylate cyclase coupled GC-A receptor, leading to the production of cyclic GMP, which, in turn, through protein kinase G, phosphorylates and activates HSL. Therefore, ANP and catecholamines converge in activating the same enzyme. At present, it is not clear why ANP- but not catecholamine-stimulated lipolysis was present in hESC-derived adipocytes. It is not due to the cell line (HS306) used in this study, because we observed similar responses in the HS346 cell line. Several reasons, including the early developmental origin of these cells, could be of importance. We have previously shown that the lipolytic response differs in adipocytes derived from developmentally different precursors (6). Nevertheless, our results show that hESC-derived adipocytes differ qualitatively from fat cells obtained from other sources. It is possible that differentiation conditions with optimal adipogenic capacity tested in this and other studies can still be improved.

Another intriguing finding from this study is the difference in differentiation between cells grown in the presence of serum and those grown in the presence of SR. Exchanging serum in standard adipogenic medium with SR resulted in an improved adipogenic effect at both the morphologic and functional level. Speculatively, this could be caused by putative anti-adipogenic factors present in FCS.

Pluripotent differentiation potential including adipogenic conversion of hESC was recently described in a set of ESC lines (9). These cells differed from those described in this study because they used a set of hESCs cultured on murine fibroblasts, whereas our cells have been established, propagated, and cultured on human fibroblasts in SR medium. Moreover, adipogenic conversion was only determined by staining with the triglyceride-specific dye Oil Red O and mRNA expression of PPAR2. In a paper by Xiong et al. (10), adipogenic conversion of hESCs including mRNA expression and leptin secretion was presented. It should be noted that neither of these studies examined mRNA expression for genes specific for adipocyte lipolysis (2-AR, 2-AR, GC-A, HSL) or lipolysis. Nevertheless, both of these studies adopted adipogenic protocols that were not too different from those used in this study. During the preparation of this manuscript, Olivier et al. (8) presented data showing that MSCs can be derived from hESCs in feeder cell-free cultures. These cells displayed osteogenic and adipogenic capacity determined by morphology and mRNA expression. No data on lipolysis or adipokine release were included.

In summary, we described lipolysis in hESC-derived adipocytes. Our main conclusion is that, although adipocyte-like characteristics of hESCs can be obtained in vitro, these cells seem to be functionally different from human mature adipocytes. Although these cells express mRNA for lipolytic genes, their responsiveness to -adrenergic agonists is very small, whereas ANP lipolysis is similar to that observed in mature adipocytes. This suggests that hESC-derived adipocytes cannot substitute for preadipocyte- or MSC-based model systems.

Notes

¹ Nonstandard abbreviations: TG, triglyceride; HSL, hormone sensitive lipase; 2A-AR, adrenergic receptor 2A; -AR, -adrenoceptor; ANP, atrial natriuretic peptide; GC-A, guanylate cyclase receptor A; MSC, mesenchymal stem cell; hESC, human embryonic stem cell; DMEM, Dulbecco's modified Eagle's medium; EB, embryonic body; F12, Ham's F12 media; FCS, fetal calf serum; SAM, standardized adipogenic medium; SR, serum replacement; RA, retinoic acid; PPAR, peroxisome proliferator-activated receptor; NA, noradrenaline; hMSC, human mesenchymal stem cell.

References

1. Arner, P. (2002) Insulin resistance in type 2 diabetes: role of fatty acids. Diabetes Metab Res Rev. 18(Suppl 2): S5–S9. | Article | PubMed | ISI | ChemPort |
2. Gimeno, R. E., Klaman, L. D. (2005) Adipose tissue as an active endocrine organ: recent advances. Curr Opin Pharmacol. 5: 122–128. | Article | PubMed | ChemPort |
3. Arner, P. (2005) Resistin: yet another adipokine tells us that men are not mice. Diabetologia 48: 2203–2205. | Article | PubMed | ISI | ChemPort |
4. Arner, P. (2005) Human fat cell lipolysis: biochemistry, regulation and clinical role. Best Pract Res Clin Endocrinol Metab. 19: 471–482. | Article | PubMed | ISI | ChemPort |
5. Sengenes, C., Bouloumie, A., Hauner, H., et al (2003) Involvement of a cGMP-dependent pathway in the natriuretic peptide-mediated hormone-sensitive lipase phosphorylation in human adipocytes. J Biol Chem. 278: 48617–48626. | Article | PubMed | ChemPort |
6. Dicker, A., Kaaman, M., van Harmelen, V., Astrom, G., Blanc, K. L., Ryden, M. (2005) Differential function of the alpha2A-adrenoceptor and Phosphodiesterase-3B in human adipocytes of different origin. Int J Obes (London) 29: 1413–1421. | Article | ChemPort |
7. Zuk, P. A., Zhu, M., Mizuno, H., et al (2001) Multilineage cells from human adipose tissue: implications for cell-based therapies. Tissue Eng. 7: 211–228. | Article | PubMed | ISI | ChemPort |
8. Olivier, E. N., Rybicki, A. C., Bouhassira, E. E. (2006) Differentiation of human embryonic stem cells into bipotent mesenchymal stem cells. Stem Cells 24: 1914–1922. | Article | PubMed | ChemPort |
9. Barberi, T., Willis, L. M., Socci, N. D., Studer, L. (2005) Derivation of multipotent mesenchymal precursors from human embryonic stem cells. PLoS Med. 2: e161 | Article | PubMed | ChemPort |
10. Xiong, C., Xie, C. Q., Zhang, L., et al (2005) Derivation of adipocytes from human embryonic stem cells. Stem Cells Dev. 14: 671–675. | Article | PubMed | ChemPort |
11. Inzunza, J., Gertow, K., Stromberg, M. A., et al (2005) Derivation of human embryonic stem cell lines in serum replacement medium using postnatal human fibroblasts as feeder cells. Stem Cells 23: 544–549. | Article | PubMed | ISI | ChemPort |
12. Hovatta, O., Mikkola, M., et al (2003) A culture system using human foreskin fibroblasts as feeder cells allows production of human embryonic stem cells. Hum Reprod. 18: 1404–1409. | Article | PubMed | ISI |
13. Dani, C., Smith, A. G., Dessolin, S., et al (1997) Differentiation of embryonic stem cells into adipocytes in vitro. J Cell Sci. 110: 1279–1285. | PubMed | ISI | ChemPort |
14. Dicker, A., Le Blanc, K., Astrom, G., et al (2005) Functional studies of mesenchymal stem cells derived from adult human adipose tissue. Exp Cell Res. 308: 283–290. | Article | PubMed | ChemPort |
15. Hoffstedt, J., Arvidsson, E., Sjolin, E., Wahlen, K., Arner, P. (2004) Adipose tissue adiponectin production and adiponectin serum concentration in human obesity and insulin resistance. J Clin Endocrinol Metab. 89: 1391–1396. | Article | PubMed | ISI | ChemPort |
16. Hellmer, J., Marcus, C., Sonnenfeld, T., Arner, P. (1992) Mechanisms for differences in lipolysis between human subcutaneous and omental fat cells. J Clin Endocrinol Metab. 75: 15–20. | Article | PubMed | ISI | ChemPort |
17. Arner, P. (1995) Techniques for the measurement of white adipose tissue metabolism: a practical guide. Int J Obes Relat Metab Disord. 19: 435–442. | PubMed | ChemPort |

Acknowledgments

This work was supported by grants from the Swedish Research Council, the Swedish Medical Association, AFA Försäkring, and the foundations of Magnus Bergvall and Söderberg.