THYROIDVolume 11, Number 3, 2001Mary Ann Liebert, Inc.

Induction of Leptin Expression in Orbital Preadipocyte Fibroblasts

Dana Z. Erickson,1 Debra A. Harteneck,1 Bradley J. Erickson,2 Charyl M. Dutton,1 and Rebecca S. Bahn1

Graves’ ophthalmopathy (GO) is an autoimmune disease characterized by an increase in the volume of the or-bital fatty/connective tissues and extraocular muscles. This volume change is due to expansion of the adiposetissues and to accumulation of glycosaminoglycans and edema within the connective tissues of the orbit. Wehave shown previously that a subpopulation of confluent human orbital preadipocyte fibroblasts can be in-duced in vitro to differentiate into cells with morphological features of adipocytes and that these cultures ex-press functional thyrotropin receptor (TSHR). In order to identify and study these cells further, we examinedthe expression of leptin protein and TSHR and leptin mRNA in these cultures. Using immunocytochemistrywith objective measurement of immunofluorescent staining intensity on digitized microscopic images, we de-termined leptin protein expression to be 6 to 37 times greater in differentiated cultures than in control cultures.In addition, we showed that the expression of both genes is enhanced in differentiated cultures. We suggestthat an unknown humoral stimulus, present in Graves’ disease, might act to induce the differentiation of nor-mal orbital fibroblasts into TSHR-bearing adipocytes. This process would be expected to result in expansion ofthe orbital adipose tissues and increased TSHR expression within the orbit.

221

Introduction

GRAVES’ OPHTHALMOPATHY (GO) is an autoimmune dis-ease characterized by an increase in the volume of adi-

pose/connective tissues and extraocular muscle bodieswithin the orbit (1). There has long been a controversy re-garding the primary autoimmune targets in this disease. Sev-eral recent studies support the concept that the thyrotropinreceptor (TSHR) is an important autoantigen in GO (2–7).The finding that mRNA corresponding to intact humanTSHR is expressed to a greater degree in orbital tissues frompatients with GO than in normal orbital tissues (3) suggeststhat a humoral factor may stimulate TSHR expression in or-bital tissues in the setting of Graves’ disease. Other studieshave described a subpopulation of orbital preadipocyte fi-broblasts that can be induced in vitro to differentiate into cellswith morphological features of adipocytes (8). The increasedexpression of functional TSHr seen in these differentiatedcultures suggests that some orbital fibroblasts may have thepotential in vivo to increase expression of this receptor, givenappropriate stimuli (4).

The present study was designed to identify the differen-tiated cells within these cultures further and to characterizethis process in an objective and quantifiable manner. Wequantitated the intensity of leptin-specific immunofluores-cent staining to determine whether differentiation of orbital

preadipocyte fibroblasts in vitro enhances expression of thisadipocyte-specific protein. In addition, we studied the ex-pression of leptin and TSHR genes using an RNase protec-tion assay designed to detect simultaneously the presence ofboth mRNA species.

Methods

Cell culture and adipocyte differentiation

Orbital adipose/connective tissue explants were obtainedfrom patients undergoing orbital decompression surgery forsevere GO. As in our previous studies (4,9), explants wereminced and placed directly in plastic culture dishes con-taining medium 199 with 20% fetal bovine serum (FBS; Hy-Clone, Logan UT), penicillin (100 U/mL), and gentamicin (20g/mL) in a humidified 5% CO2 incubator at 37°C. For use inimmunohistochemical studies, adherent fibroblasts wereharvested, seeded onto glass chamber slides (Nalge Nunc In-ternational, Rochester, NY), and grown to confluence asmonolayers in medium 199 containing 10% FBS. For cellulardifferentiation, monolayer cultures were switched to serum-free Dulbecco’s Modified Eagle’s Medium/F12 (1:1; Sigma,St. Louis, MO) supplemented with insulin (1 M), tri-iodothyronine (0.2 nM), carbaprostacyclin (cPGI2; 0.2 M; Cal-biochem, La Jolla, CA), biotin (33 M), pantothenic acid (17

1Division of Endocrinology, Metabolism and Nutrition, and 2Department of Diagnostic Radiology, Mayo Clinic/Foundation, Rochester,Minnesota.

M), transferrin (10 g/mL) and for the first 4 days only, dex-amethasone (1 M) and isobutylmethylxanthine (IBMX; 0.1mM). Medium was replaced every 3–4 days of culture. Touse as control cells, fibroblasts derived from the same pa-tients’ orbital tissues were cultured similarly except for theomission in the medium of cPGI2, dexamethasone and IBMX.Each slide chamber contained approximately the same num-ber of cells (1 3 105) at the end of the culture period. For usein RNase protection assays, cells were passaged and grownto confluence in medium 199 with 10% FBS and then eithersubjected to the differentiation protocol as above, grown ascontrol cultures in the same medium lacking cPGI2, dexa-methasone and IBMX, or maintained in medium 199 with10% FBS.

Immunofluorescent staining for leptin

Monolayer cells on glass chamber slides (n 5 4 patients)were chilled on ice, the media were removed, and the cellswere washed with Tris-buffered saline (TBS). Slides were in-cubated with sodium bromide (0.5%; Sigma S-9125) for 5minutes to decrease autofluorescence, rewashed, and fixedin methanol for 15 minutes at 220°C. Nonspecific stainingwas blocked with goat serum (0.5%) for 30 minutes. Poly-clonal antileptin rabbit antibody directed against amino acid91–106 of human leptin (1:500 dilution; Affinity BioReagents,Inc., Golden, CO) was applied for 2 hours at room temper-ature in a humid chamber. Slides were washed three timeswith TBS-Tween 20 and incubated in the dark at room tem-perature for 30 minutes with secondary anti-rabbit Texas Redantibody (1:350 dilution; Vector Laboratories, Burlingame,CA). Slides were again washed three times. After applica-tion of Vectashield mounting medium (Vector Laboratories;TI-1000; Burlingame, CA), coverslips were applied and cellswere examined immediately under a LSM 510 confocal flu-orescent microscope (Carl Zeiss, Inc., Oberkochen, Germany)at excitation 568 nm. Parallel slides with primary or sec-ondary antibodies replaced in turn by TBS-Tween 20 ormatched nonimmune rabbit immunoglobulin G (IgG) wereprocessed to ensure specificity and exclude cross-reactivitiesbetween the tagged antibodies used.

Quantitative analysis of leptin-specific staining intensity

Two to four random photographic images of each glasschamber slide were obtained using uniform optics and dig-itization (n 5 3 patients). The objective used was an C-Apo-chromat 403/1.2 water immersion lens. Quantitation of theintensity of immunofluorescence was performed on digi-tized microscope images using the ANALYZE image analy-sis package developed at the Mayo Clinic (10). Measure-ments of background intensity and variance indicated thatnoncell pixels were consistently below 60 arbitrary units(AU). Therefore, the number of pixels showing intensitygreater than 60 AU was considered to represent specific im-munofluorescent staining. Image analysis was performed ina blinded fashion.

RNAse protection assay (RPA) for TSHr and leptin gene expression

After experiments in vitro, cells were pelleted and storedat 270°C until processed for RNA isolation. Total RNA was

isolated directly from specimens using the Totally RNA Kit(Ambion, Austin, TX). Positive control RNA was preparedin the same manner from normal human thyroid tissue (forTSHR) and human abdominal fat (for leptin). For use in gels,5 g of each was combined and run as a single positive con-trol lane. Negative control RNA was prepared from Epstein-Barr virus (EBV)-transformed human B cells.

The antisense RNA probe for TSHR was transcribed froma 320 base pair (bp) polymerase chain reaction (PCR) prod-uct with a T7 phage promoter at its 39 end, in the presenceof T7 RNA polymerase (10 units) and [32P]UTP (50 Ci) forlabeling (3,4). The resulting high specific-activity probe en-compassed nucleotides 576–873 (exons 6–9) of the humanTSHR cDNA sequence and was designed to detect both the2.4-kb intact TSHR (protecting a product of 298 nucleotides)and the 1.3-kb variant form (protecting a product of 217 nu-cleotides). The antisense RNA probe for human leptin wastranscribed from a 201-bp PCR product encompassing nu-cleotides 185–357 (exon 3) of the human leptin cDNA se-quence (11). This probe was designed to protect a 172-nu-cleotide fragment of human leptin mRNA. The antisenseRNA probe for glyceraldehyde 3-phosphate dehydrogenase(GAPDH) was generated from pTRI-GAPDH human anti-sense control template (Ambion, Austin, TX). This probe wasdesigned to protect a 154-nucleotide fragment of GAPDHmRNA.

Total RNA (80 mg) was combined with 300,000 cpm TSHRand leptin probes and 3,000 cpm (GAPDH probe in hy-bridization buffer, denatured at 95°C, and hybridized at 45°Cfor 16 hours. Nonhybridized total RNA and probes were di-gested for 1 hour at 37°C with RNAase A (0.175 units) andRNAase T1 (25 units; RNAase Protection Kit, BoehringerMannheim, Indianapolis, IN). Samples were subsequentlydigested with proteinase K (50 g) in the presence of 0.5%sodium dodecyl sulfate (SD), and extracted with phenol/chloroform/isoamyl alcohol. The resulting ethanol-precipi-tated protected fragments were resuspended in loadingbuffer and resolved on a denaturing polyacrylamide gel (5%acrylamide/8 M urea).

Results

Leptin-specific immunofluorescent staining

Specific staining for leptin protein first became evident be-tween days 1 and 5 of differentiation culture, and peaked bydays 7 to 9. At this point, approximately 10%–20% of the cellswithin each culture (n 5 4) showed bright perinuclear andcytoplasmic immunofluorescent staining for leptin (Fig. 1A).The cells in culture were morphologically heterogeneous;some appeared elongated and fibroblast-like, while otherswere larger and more rounded. Leptin-specific staining ap-peared to be present primarily in the larger of the culturedcells and was significantly less apparent by day 12 of differ-entiation than it had been at day 7. In contrast, GO fibro-blasts from the same patients cultured in control medium199 with 10% FBS revealed only minimal leptin-specificstaining through the entire 12 day period (Fig. 1B). Neitherslides of differentiated cells incubated with nonimmune rab-bit IgG (Fig. 1C), nor slides with primary or secondary anti-bodies replaced in turn by TBS-Tween 20 (not shown)showed significant immunofluorescent staining.

ERICKSON ET AL.222

Quantitative analysis of leptin-specific staining

Computer-assisted quantitative analysis of the intensity ofleptin-specific immunofluorescence (n 5 3) revealed that themean number of pixels with fluorescence measured at morethan 60 AU (considered to represent background fluores-cence) was significantly greater in photomicrographs of dif-ferentiated day 7 cultures (mean 5 3969 pixels) than in con-trol matched cultures (mean 5 133 pixels; Fig. 2). The widerange in number of positive pixels in the individual differ-entiated (range, 220–13,854 pixels) or control (range, 22–362pixels) cultures reflects our observation that there was fairlymarked patient-to-patient variation in the magnitude of lep-tin staining. However, when photomicrographs of each pa-tient’s differentiated day 7 cells were compared to images ofthat same patient’s cells maintained in control 199 medium,the former were found to have between 16 and 37 times asmany pixels over 60 AU than did the latter (Fig. 2). On day12, leptin staining was considerably less intense in all dif-ferentiated cultures (mean pixels over 60 AU 5 259) com-

pared with day 7, and the difference in mean staining in-tensity between patient-matched differentiated and controlcultures was only approximately 1.3-fold (data not shown).In differentiated cultures stained with isotype-matched ornonimmune rabbit IgG in place of anti-rabbit leptin anti-body, the number of pixels having immunofluorescencegreater than 60 AU was consistently very low (mean 14 pix-els).

Expression of TSHr and leptin genes

Using an RPA designed to detect simultaneously the ex-pression of both TSHR and leptin genes in orbital cells (n 53), we studied cultures maintained for 8 days in adipocytedifferentiation medium. Control cultures for these experi-ments were maintained for 8 days in the same medium lack-ing cPGI2, dexamethasone and IBMX. Leptin mRNA wasclearly detected in differentiated cultures, while little or noevidence of leptin gene expression was evident in these con-trol cultures (Fig. 3) or in control cultures maintained in

INDUCTION OF LEPTIN EXPRESSION 223

A B

C

FIG. 1. A: Leptin-specific immunofluorescent stainingof differentiated orbital preadipocyte fibroblasts from aGraves’ ophthalmopathy (GO) patient demonstratingbright perinuclear and cytoplasmic staining on day 7 ofdifferentiation. B: Orbital preadipocyte fibroblasts fromthe same GO patient cultured in control medium 199with 10% fetal bovine serum (FBS) revealing minimalleptin-specific immunofluorescent staining. C: Differen-tiated orbital cells incubated with isotype control non-immune immunoglobulin G (IgG) in place of antileptinprimary antibody showing only background fluores-cence. Experiments were performed using cells fromthree additional GO patients with similar results.

medium 199 with 10% FBS (n 5 2; data not shown). In con-trast, TSHR gene expression was clearly apparent both in dif-ferentiated and in control cultures lacking cPGI2, dexa-methasone and IBMX, while only faint evidence of TSHRmRNA was seen in control cultures maintained in medium199 with 10% FBS.

Discussion

The increased orbital tissue volume characteristic of GOis attributable both to expansion of the adipose tissues andto accumulation of hydrated GAG within the connective tis-sues and extraocular muscles (12). The latter typically resultsin edematous enlargement of one or more of the extraocularmuscles and is generally accompanied by an increase in or-bital adipose tissue volume (13). This increase is not likelydue to edema alone, as the density of orbital fat tissue (afunction of the state of hydration) seen in computerized to-mography scans of GO patients is similar to that of normalorbital fat (14). Postmortem examination of GO orbital adi-pose tissues demonstrates that both the weight and volumeare increased compared to normal tissues, and that thesevariables correlate with the degree of exophthalmos present(15).

Adipocyte precursor cells, called preadipocytes, can beisolated from the stromal-vascular fraction of neonatal andadult human adipose/connective tissues from various re-gions of the body (16). These cells appear to belong to a sub-population of fibroblasts having the potential to undergoadipocyte differentiation when exposed to appropriate invitro conditions (17). A report by Sorisky and colleagues (8)demonstrated that cultures derived from orbital connectivetissue contain such adipocyte precursor cells (comprising5%–10% of the total), capable of responding to adipogenicstimuli.

During adipogenesis, a programmed induction of variousgenes regulates lipoprotein lipolysis, the cellular uptake of

fatty acids, and the synthesis of fatty acids and triglycerides(18,19). Leptin, the product of the obese (ob) gene, is expresseduniquely in cells committed to the adipocyte lineage and isexpressed during late or terminal differentiation (20,21).Thus, this protein can serve as a useful marker of adipocytedifferentiation. In order to better understand the phenotypicchanges observed in orbital preadipocyte fibroblasts duringdifferentiation, we studied the expression of this protein overtime in cultures of cells from patients with GO. Using im-munohistochemistry, we determined that this protein firstbecomes evident in a subpopulation of cells (perhaps 10%)between days 1 and 5 of differentiation culture, peaking be-tween days 7 and 9. This profile of leptin expression is some-what different from that seen in the immortalized murine3T3-L1 cells, a preadipocyte cell line that has also been shownto express functional TSHR upon differentiation (22,23). In3T3-L1 cultures, greater than 90% of cells differentiate intomature adipocytes by day 4 (24). Perhaps paracrine factorssecreted by differentiating cells speed the rate of differenti-ation. If so, the relatively slow rate of adipogenesis seen inthe orbital cultures might reflect the fact that only about 10%of cells in these inhomogeneous cultures undergo differen-tiation.

ERICKSON ET AL.224

FIG. 2. Quantitative analysis of leptin-specific immunoflu-orescent staining on day 7 in differentiated (open bars) andmatched control (solid bars) cultures maintained in medium199 with 10% fetal bovine serum (FBS) (n 5 3). The abscissarepresents the number of pixels having fluorescence value.60 arbitrary units (AU) the value considered to representbackground fluorescence. Each bar represents the mean ofthe analysis of 3–4 different digitized microscopic images.

FIG. 3. RNase protection assay of thyrotropin receptor(TSHR) and leptin mRNA. Lane 1: Molecular size standardsranging from 100 to 500 bp. Lane 2: Negative control humanEBV-transformed B cells. Lane 3: Positive control combinedhuman thyroid and fat tissues. Lane 4: Differentiated fi-broblasts from patient with Graves’ ophthalmopathy (GO).Lane 5: Control fibroblasts from same patient with GO maintained in the same medium lacking carbaprosta-cyclin (cPGI2), dexamethasone, and isobutylmethylxanthine(IBMX). Positive protected bands at 298 bp and 217 bp (seenin lanes 3–5) correspond to TSHR mRNA. Leptin mRNA isrepresented by bands at 172 bp (seen in lanes 3 and 4), whileglyceraldehyde-3-phosphate dehydrogenase (GAPDH) bandsare apparent at 154 bp in lanes 2–5. Similar results werefound using cells from two additional GO patients.

Because precise quantitation of leptin staining is not pos-sible using standard immunohistochemical techniques, wedeveloped a method by which staining intensity could becaptured and measured on digitized microscopic images.This technique revealed a 17- to 36-fold increase in the quan-tity of leptin present in differentiated cultures on day 7 com-pared with control cultures. Because the expression of lep-tin appears to develop concomitant with morphologicalchanges in a subset of the cells (8), it is likely that these cellsare indeed newly differentiated adipocytes. It is unclear whyleptin expression peaks around day 7 in culture and becomesless apparent by day 12. This feature may be a function ofthe culture environment itself, with the cells themselves per-haps becoming less viable after a prolonged period in cul-ture.

These experiments were designed, in part, to study thepossible linkage between adipogenesis (as marked by leptingene expression) and our previous finding of TSHR gene ex-pression in orbital preadipocyte fibroblasts (4). While wedemonstrated the presence of leptin mRNA only in differ-entiated cultures, we showed clear evidence of TSHR mRNAboth in differentiated cultures and in control cultures lack-ing cPGI2, dexamethasone, and IBMX. In contrast, nostrongly positive bands corresponding to TSHR or leptinmRNA were apparent in control cultures maintained inmedium 199 with 10% FBS. Because these two “control” me-dia appear to have dissimilar effects on TSHR and leptin ex-pression in these cells, we suggest that TSHR and leptingenes may be independently expressed in these cultures.These studies using RPA are limited by their inability to de-termine precisely which cells contain TSHR or leptin mRNA.Therefore, it is unclear whether these genes are expressed inthe same or different subpopulations within these heteroge-neous cultures. Alternately, it may be that both genes are in-deed expressed in the same cells, but that expression of eachis stimulated by somewhat different factors present in theculture conditions used. Our immunohistochemical studiesshowed that leptin protein is present essentially only in thatsubpopulation of cells with morphological features ofadipocytes. Clearly, simultaneously detection of both leptinand TSHr using immunohistochemistry or in situ hybridiza-tion would help to clarify these issues.

The orbital cells used in these studies were obtained frompatients with GO, rather than from normal individuals.However, we have experience in culturing normal orbitalcells under these same adipogenesis-stimulating conditionsand observe similar morphological changes in these normalcultures. Furthermore, we showed previously that normalorbital fibroblasts, similar to Graves’ orbital fibroblasts, in-crease expression of TSHR when cultured under these con-ditions (4). Therefore, we expect that normal orbitalpreadipocyte fibroblasts, like Graves’ fibroblasts, would alsoincrease leptin gene expression in differentiation culture.Our hypothesis is that the disease-specificity resides not inthe orbital cells themselves, but in a yet to be defined dis-ease-specific circulating factor(s) present in patients withGraves’ disease.

In summary, a subpopulation of orbital fibroblasts has theability to undergo in vitro differentiation into leptin-ex-pressing adipocytes. Additionally, we reported previouslythat TSHR gene expression is enhanced during differentia-tion in these cultures (4). In the setting of Graves’ disease,

we suggest that an unknown disease-specific factor(s) stim-ulates both the differentiation of orbital preadipocyte fi-broblasts into adipocytes and increased expression of TSHRwithin the orbit, perhaps within the same cell subpopula-tion. This process would be expected to lead to expansion ofthe orbital adipose tissue volume, resulting in many of thecharacteristic clinical signs and symptoms of the disease. Inaddition, the accompanying increase in expression of the pu-tative autoantigen, TSHR, might play an important role inpropagation of the orbital autoimmune response. Futurestudies will be aimed at defining the factors responsible forthis orbital process in patients with GO.

Acknowledgments

We would like to thank Mr. James Tarara for his assistancewith the immunofluorescent imaging.

This work was supported in part by NIH EYO8819 (toR.S.B.).

References

1. Bahn RS, Heufelder AE 1993 Pathogenesis of Graves’ oph-thalmopathy. N Engl J Med 329:1468–1475.

2. Many M-C, Costagliola S, Detrait M, Denef F, Vassart G,Ludgater MC 1999 Development of an animal model of au-toimmune thyroid eye disease. J Immunol 162:4966.

3. Bahn RS, Dutton CM, Natt N, Joba W, Spitzweg C, HeufelderAE 1998 Thyrotropin receptor expression in Graves’ orbitaladipose/connective tissues: Potential autoantigen in Graves’ophthalmopathy. J Clin Endocrinol Metab 83:998–1002.

4. Valyasevi RW, Erickson DZ, Harteneck DA, Dutton CM,Heufelder AE, Jyonouchi SC, Bahn RS 1999 Differentiationof human orbital preadipocyte fibroblasts induces expres-sion of functional thyrotropin receptor. J Clin EndocrinolMetab 84:2557–2562.

5. Perros P, Kendall-Taylor P 1994 Demonstration of thy-rotropin binding sites in orbital connective tissue: Possiblerole in the pathogenesis of thyroid-associated ophthal-mopathy. J Endocrinol Invest 17:163–170.

6. Mengistu M, Lukes YG, Nagy EV, Burch HB, Carr FE, LahiriS, Burman KD 1994 TSH receptor expression in retroocularfibroblasts. J Endocrinol Invest 17:437–441.

7. Feliciello A, Porcellini A, Ciullo, Bonavolonta G, Avvedi-mento EV, Fenzi G 1993 Expression of thyrotropin-receptormRNA in healthy and Graves’ retro-orbital tissue. Lancet342:337–338.

8. Sorisky A, Pardasani D, Gagnon A, Smith TJ 1996 Evidenceof adipocyte differentiation in human orbital fibroblasts inprimary culture. J Clin Endocrinol Metab 81:3428–3431.

9. Bahn RS, Gorman CA, Woloschak GE, David CS, JohnsonPM, Johnson CM 1987 Human retroocular fibroblasts invitro: A model for the study of Graves’ ophthalmopathy. JClin Endocrinol Metab 65:665–670.

10. Robb RA, Hanson DP 1996 The ANALYZE software systemfor visualization and analysis in surgery simulation. In: Tay-lor RH, LaVallee S, Burdea GC, Mosges R, eds. ComputerIntegrated Surgery. MIT Press, Cambridge, MA, pp 175–189.

11. Masuzaki H, Ogawa Y, Isse N, Satoh N, Okazaki T, Shige-moto M, Mori K, Tamura N, Hasada K, Yoshimasa Y 1995Human obese gene expression. Adipocyte-specific expres-sion and regional differences in the adipose tissue. Diabetes44:855–858.

12. Roselli-Rehfuss L, Robbins LS, Cone RD, Hufnagel TJ,

INDUCTION OF LEPTIN EXPRESSION 225

Hickey WJ, Cobbs WH, Jacobiec FA, Iwamoto T, Eagle RC1987 Immunohistochemical and ultrastructural studies onthe exenterated orbital tissues of a patient with Graves’ dis-ease. Ophthalmopathy 91:1411–1419.

13. Forbes G, Gorman CA, Brennan MD, Gehring DG, IlstrupDM, Earnest F 1986 Ophthalmopathy of Graves’ disease:computerized volume measurements of the orbital fat andmuscle. American Journal of Neuroradiology 7:651–656.

14. Peyster RG, Ginsberg F, Silber JH, Adler LP 1986 Exoph-thalmos caused by excessive fat: CT volumetric analysis anddifferential diagnosis. American Journal of Neuroradiology7:35–40.

15. Rundle FF, Pochin EE 1973 The orbital tissues in thyrotoxi-cosis: a quantitative analysis relating to exophthalmos. ClinSci 5:51–74.

16. Poznanski WJ, Waheed I, Van R 1973 Human fat cell pre-cursors. Morphologic and metabolic differentiation in cul-ture. Lab Invest 29:570–576.

17. Hausman GJ, Richardson RL 1998 Newly recruited and pre-existing preadipocytes in cultures of porcine stromal-vascu-lar cells: Morphology, expression of extracellular matrixcomponents, and lipid accretion. J Anim Sci 76:48–60.

18. Sorisky A 1999 From preadipocyte to adipocyte: Differenti-ation-directed signals of insulin from the cell surface to thenucleus. Crit Rev Clin Lab Sci 36:1–34.

19. Smas CM, Sul HS 1995 Control of adipocyte differentiation.Biochem J 309:697–710.

20. Mitchell SE, Rees WD, Hardie LJ, et al 1997 Ob gene ex-

pression and secretion of leptin following differentiation ofrat preadipocytes to adipocytes in primary cultures.Biochem Biophys Res Commun 230:360–364.

21. Leroy P, Dessolin S, Villageios P, Moon BC, Friedman JM,Ailhaud G, Dani C 1996 Expression of the ob gene in adi-pose cells: Regulation by insulin. J Biol Chem 27:2365–2368.

22. Haraguchi K, Shimura H, Lin L, Saito T, Endo T, Onaya T1996 Functional expression of thyrotropin receptor in dif-ferentiated 3T3-L1 cells: a possible model cell line of ex-trathyroidal expression of thyrotropin receptor. BiochemBiophys Res Commun 223:193–198.

23. Shimura H, Haraguchi K, Endo T, Onaya T 1997 Regulationof thyrotropin receptor gene expression in 3T3-L1 cells isdistinct from its regulation in FRTL-5 thyroid cells. En-docrinology 138:1483–1490.

24. Cowherd RM, Lyle RE, McGehee RE 1999 Molecular regu-lation of adipocyte differentiation. Semin Cell Dev Biol10:3–10.

Address reprint requests to:Rebecca S. Bahn, M.D.

Mayo ClinicDivision of Endocrinology

200 First Street, SWRochester, MN 55905

E-mail: bahn.rebecca@mayo.edu

ERICKSON ET AL.226