J. Clin. Endocrinol. Metab. 2001 86: 4920-4925, doi: 10.1210/jc.86.10.4920  

F. Savagner, B. Franc, S. Guyetant, P. Rodien, P. Reynier and Y. Malthiery  

Defective Mitochondrial ATP Synthesis in Oxyphilic Thyroid Tumors

Society please go to: http://jcem.endojournals.org//subscriptions/ or any of the other journals published by The EndocrineJournal of Clinical Endocrinology & Metabolism To subscribe to

Copyright © The Endocrine Society. All rights reserved. Print ISSN: 0021-972X. Online

Defective Mitochondrial ATP Synthesis in OxyphilicThyroid Tumors

F. SAVAGNER, B. FRANC, S. GUYETANT, P. RODIEN, P. REYNIER, AND Y. MALTHIERY

Inserm EMI-U 00-18 (F.S., P.R., P.R., Y.M.), Laboratoire de Biochimie et Biologie Moleculaire, Angers F-49033; Laboratoired’Anatomie Pathologique (B.F.), Hopital Ambroise Pare, Boulogne F-92104; Laboratoire d’Anatomie Pathologique (S.G.);and Service d’Endocrinologie (P.R.), Nutrition et Medecine Interne, Angers F-49033, France

Oxyphilic tumors (oncocytomas or Hurthle cell tumors) forma rare subgroup of thyroid tumors characterized by cells con-taining abundant mitochondria. The relationship betweenthe mitochondrial proliferation and the pathogenesis of thesetumors is unknown. We have assessed the expression of themitochondrial ND2 and ND5 (subunits of the nicotinamideadenine dinucleotide dehydrogenase complex) genes and thenuclear UCP2 (uncoupling protein 2) gene in 22 oxyphilicthyroid tumors and matched controls. The consumption of

oxygen in mitochondria from tumors was determined by po-larography. ATP assays were used to explore the mitochon-drial respiratory chain activity and the oxidative phosphor-ylation coupling in seven fresh thyroid tumors and controls.Adenosine triphosphate synthesis was significantly lower inall the tumors, compared with controls, suggesting that a cou-pling defect in oxidative phosphorylation may be a cause ofmitochondrial hyperplasia in oxyphilic thyroid tumors. (JClin Endocrinol Metab 86: 4920–4925, 2001)

OXYPHILIC THYROID TUMORS, also known as onco-cytomas or Hurthle cell tumors, represent a rare sub-

group of follicular thyroid neoplasms, characterized by cellswith a distinctive eosinophilic cytoplasm (1). The cytoplas-mic eosinophilia is owing to the abundance of morpholog-ically altered mitochondria in the majority of tumor cells (1,2). However, the relationship between mitochondrial prolif-eration and the histogenesis of oxyphilic tumors is unknown.

The development of mitochondria involves the synthesisof proteins encoded by mitochondrial DNA (mtDNA), whichcarries the genes for the essential subunits of the respiratorychain complexes, as well as by nuclear DNA (nDNA). Elec-trons from the nicotinamide adenine dinucleotide generatedby glycolysis in the cell are transported into the mitochondriain which the flow of electrons between the respiratory chaincomplexes supplies the energy used by ATP synthase toproduce ATP. An alternative source of energy, independentof ATP synthase, is provided by the uncoupling protein(UCP), which plays an important role in energy homeostasis(3). The analysis of mtDNA in several oxyphilic tumors hasshown that the abnormally high histochemical activity of therespiratory chain complexes is associated with a great in-crease in the amount of wild-type mtDNA (4).

The frequency of aneuploid or polyploid cells in oxyphilicthyroid tumors suggests the presence of anomalies in nucleargenes (5). Mitochondrial proliferation has also been reportedin mitochondrial diseases associated with respiratory chaindefects or coupling defects between the respiratory chain andATP production (6–9). Among the UCPs that induce ther-mogenesis during mitochondrial respiration (3), UCP2 is theunique form expressed in many tissues and cell types. Inparticular, UCP2 is expressed in thyroid tissue, whereas theexpression of UCP1 and UCP3 is limited to brown fat adipose

tissue and muscle/white fat tissues, respectively (10). How-ever, the histochemical investigation of the key nuclear com-ponents involved in the oxidative phosphorylation processhas revealed no coupling defects in oxyphilic thyroid tumors(11).

The proliferation of mitochondria in oxyphilic tumorsmight result from the induction of genes involved in mito-chondrial biogenesis. The patterns of mitochondrial tran-scripts (especially ND2 and ND5) as well as nuclear tran-scripts from certain genes coding for proteins involved inoxidative phosphorylation differ according to whether thetumor is a renal oxyphilic tumor or a salivary gland oxyphilictumor (12). Changes in mtDNA transcription and mitochon-drial mRNA stability were observed in the former but not inthe latter, suggesting that the process of mitochondrial pro-liferation varies according to the origin of the tumor.

We examined thyroidectomy specimens from 22 anony-mous patients. In each case, tissue removed from the normalpart of the thyroid gland served as a control for the oxyphilicthyroid tumor. For each set of paired specimens, we deter-mined the gene expression profiles of the mitochondrial ND2and ND5 genes as well as the nuclear UCP2 gene involvedin energy production. Fresh tissue samples, obtained from 7of the 22 patients, were analyzed by polarography to inves-tigate the mitochondrial respiratory chain activity, the rate ofoxidative phosphorylation (ADP/oxidation ratio) and ATPassays were used to determine the mitochondrial ATPsynthesis.

Materials and MethodsThyroid tissue samples

Twenty-two benign or malignant oxyphilic thyroid tumors, diag-nosed between 1992 and 2000 at the Ambroise Pare Hospital, Paris (15cases) and the University Hospital, Angers (7 cases), were included inthe study. All the samples used were rendered anonymous (i.e. allpatient identifiers were deleted before the study). The cases were con-

Abbreviations: mtDNA, Mitochondrial DNA; nDNA, nuclear DNA;rDNA, ribosomal DNA; UCP, uncoupling protein.

0013-7227/01/$03.00/0 The Journal of Clinical Endocrinology & Metabolism 86(10):4920–4925Printed in U.S.A. Copyright © 2001 by The Endocrine Society

4920

secutive and unselected apart from exclusions on account of insufficientmaterial or association of the tumors with chronic thyroiditis. Nineteenof the tumors were follicular oxyphilic adenomas (six of which weretrabecular) and three were follicular oxyphilic carcinomas. Five of theadenomas were associated with a multinodular goiter. The diagnoseswere made according to the World Health Organization classification(1). Oxyphilic adenoma was distinguished from carcinoma on the basisof vascular or capsular invasion or metastasis. The patients were 2 menand 20 women, with a mean age of 53 yr (range 27–82 yr).

The average size of the tumor was 37.8 � 20.5 mm (mean � sd; range15–90 mm). In addition to neoplastic thyroid samples, normal thyroidsamples were taken sufficiently distant from the tumors to serve ascontrols. All the samples were immediately stored in liquid nitrogenuntil extraction of high-molecular-weight DNA and RNA.

The 22 tumor samples and controls were fixed in formalin, embeddedin paraffin, and stained with hematoxylin and eosin. Immunohisto-chemistry was performed on paraffin-embedded sections. Tissue sam-ples from the seven fresh tumors (six adenomas and one carcinoma) andmatched controls were kept in a preservative medium [100 mm sucrose,1 mm EGTA, 20 mm 3-[N-morpholino]propanesulfonic acid (pH 7.4), 1g/liter bovine albumin] to prepare the mitochondria for polarographicstudies and ATP measurement.

Tumoral and control tissues were compared using a nonparametrictest for matched-pair samples (Wilcoxon test), the differences beingconsidered statistically significant at P less than or equal to 0.05. All thenumerical values below are expressed as means � sd.

Immunohistochemistry

After morphological examination of hematoxylin- and eosin-stainedsections, corresponding 3-�m sections of the paraffin blocks were pre-pared for the detection of mitochondrial antigen expression as a semi-quantitative index of mitochondrial biogenesis. A monoclonal antibody113–1 was used, recognizing an unknown 60-kDa nonglycosylated pro-tein component of human cell mitochondria (BioGenex Laboratories,Inc., San Ramon, CA). Immunostaining was performed with the stan-dard avidin-biotin peroxidase technique with antigen retrieval. For neg-ative control slides, the primary antibody was either omitted or replacedby a suitable concentration of normal IgG of the same species.

Polarographic studies

Oxygen consumption was measured with a Clark electrode at 30 Cin a 2-mL chamber (Oxygraph OROBOROS, Anton Paar, Innsbruck,Austria). The chamber was isolated from contact with the atmosphereby a close-fitting cap so that the electrode current was proportional tothe partial pressure of oxygen in the sample. Substrates for the differentcomplexes of the respiratory chain were introduced into the chamberand consumed by mitochondria in the oxyphilic tumors or matchedcontrols. Mitochondria were isolated from seven fresh tissue samplesand matched controls using the standard procedure (13). Oxygen uptakewas measured after the addition of mitochondria (200 �g protein) to 2ml incubation medium [300 mm mannitol, 10 mm KH2PO4 (pH 7.2), 10mm KCl, and 5 mm MgCl2] to determine the basal respiratory activity.Substrates and inhibitors were introduced into the oxygraph chamberthrough the stopper port. First, 2 �l EDTA (20 mm) and 2 �l rotenone(5 mm) were added to inhibit exogenous ATPase and complex I (nico-tinamide adenine dinucleotide ubiquinone oxidoreductase of the respi-ratory chain), respectively. Then, 20 �l succinate (1 m) and 10 �l ADP(30 mm) were successively added to the sample to determine the rate ofoxidative phosphorylation (ADP/O ratio). Finally, 10 �l potassium cy-anide (200 mm) were added to stop oxygen uptake by inhibiting complexIV (cytochrome c oxidase of the respiratory chain).

ATP measurement

Mitochondrial ATP was measured by bioluminescence using theluciferin-luciferase reaction (Enliten, Promega Corp., Madison, WI) (14).Mitochondria were isolated, using the same methods as for the polaro-graphic studies, from the seven fresh oxyphilic tumor samples and theirmatched controls. After incubation of mitochondria with 10 mm gluta-mate and malate for 10 min, the rate of ATP synthesis (expressed per

milligram of mitochondrial protein) was determined for intact and per-meabilized mitochondria.

DNA isolation and Southern blot analysis

DNA was isolated using the phenol-chloroform procedure. The sam-ples were digested overnight with RNase A (20 �g/ml) and proteinaseK (20 mg/ml) at 37 C in Tris-HCl 10 mm, EDTA (pH 8) 0.1 m, and SDS0.5%. The proteins were removed by organic extraction followed byethanol precipitation with NaCl 0.2 m and centrifugation for 15 min at10,000 g. Five micrograms of DNA were digested with the restrictionenzyme XbaI (Biolabs, Beverly, MA). Southern blotting was performedaccording to standard methods. Probes labeled by digoxigenin wereobtained by multirandom priming and were revealed with antidigoxi-genin antibodies labeled by alkaline phosphatase (DigDNA labeling anddetection kit, Roche, Basel, Switzerland). The mitochondrial and nDNAwere detected by using the probes 12S ribosomal DNA (rDNA) (nt592-1344) and 18S rDNA (nt 1201–1811), respectively. For each sample,the intensities of the mtDNA signals, and the corresponding nDNAsignals, were quantified by densitometric analysis (Molecular Analyst,Bio-Rad Laboratories, Inc., Cambridge, MA).

RNA isolation and cDNA synthesis

RNA was isolated using the guanidinium isothiocyanate procedure(Trizol Reagent, Life Technologies, Inc., Gaithersburg, MD). ResidualDNA was removed by DNase treatment: 5 �g total RNA were incubatedwith 2 U RNase-free DNase I for 1 h at 37 C.

To generate cDNA, 1 �g of RNA was first denatured at 70 C with 1�m of oligodT (Promega Corp.) for 5 min before quenching on ice; then0.5 mm of each of the 4 dNTPs, 10 mm dithiothreitol, 10 U RNaseinhibitor, and 200 U superscript II (Life Technologies, Inc.) were addedto the 5� buffer to make up a final volume of 20 �l reaction mix. Thereaction mix was incubated for 1 h at 42 C. The reverse transcriptase wasinactivated at 70 C for 15 min.

Quantitative PCR analysis

Real-time quantitative PCR with an external standard was used todetermine the gene copy number (Lightcycler, Roche). Standard PCRproducts for each gene were generated by amplifying nuclear cDNA ormtDNA templates. PCR products were purified by the phenol-chloro-form method and the copy number in the final sample was determinedby two independent methods (i.e. spectrophotometry and gel analysis).For each gene tested, a sequence-specific standard curve was plottedusing serial dilutions of the target gene standard PCR product, and thesame primers were used to amplify the cDNA.

The expression of two mitochondrial genes, ND2 and ND5, and twonuclear genes, UCP2 and �-ACTIN, was analyzed using the PCR primersets indicated in Table 1. The amount of RNA determined for eachsample was normalized by the quantification of the �-ACTIN transcripts.

Two microliters of master mix containing Taq DNA polymerase,dNTPs, and SYBR green I (DNA Master SYBR Green I kit, Roche) wereincubated for 5 min at room temperature with 0.16 �l of Taqstart an-tibody. The PCR reaction was then started by adding MgCl2 4 mm andforward and reverse primers (0.5 �m) to the capillary tubes of theLightcycler apparatus containing the master mix and 2 �L of template(cDNA or a standard with a known copy number) in a final volume of20 �l.

TABLE 1. Oligonucleotide pairs used for quantitative PCR

Primer pairs Genes

5�-GCACCCCTCTGACATCC-3� ND25�-CGGTCGGCGAACATCAGTGG-3�5�-GGGGATTGTGCGGTGTGTG-3� ND55�-CTTCTCCTATTTATGGGGGT-3�5�-CCAGTGCGCGCGCTACAGTCA-3� UCP25�-GTGGTGCTGCCTGCTAGGAG-3�5�-CGACATGGAGAAAATCTGGC-3� �-ACTIN5�-AGGTCCAAGACGCAGGATGG-3�

Savagner et al. • ATP Synthesis in Oxyphilic Thyroid Tumors J Clin Endocrinol Metab, October 2001, 86(10):4920–4925 4921

ResultsImmunohistochemistry

The 60-kDa mitochondrial protein was present in the cy-toplasm of all the 22 oxyphilic thyroid tumor samples, bothhomogeneous (19 tumors) and heterogeneous (3 tumors)distributed (histochemistry score data not shown), confirm-ing the increased mitochondrial biogenesis in the tumor sam-ples. In contrast, the immunostaining of the matched controlsamples was extremely weak or undetectable. The homoge-neous distribution of mitochondrial immunostaining withno difference between adenomas vs. carcinomas has beenpreviously described for oxyphilic tumors (15).

Polarographic analysis

Seven tumors and matched controls (n � 14) were used forthe functional analysis of the respiratory chain. The mito-chondria were prepared within 1 h after thyroidectomy so asto obtain interpretable results. Polarographic analysis ofcomplexes I to IV produced no evidence of respiratory chaindefects in the seven oxyphilic tumors vs. control tissue. Therespiratory control indices were calculated by dividing therate of oxygen uptake in state 3 (stimulated by ADP addition)by the rate of oxygen uptake in state 4 (when all the ADP isconverted to ATP). Using succinate as substrate, the indiceswere 3.8 � 0.8 for tumors and controls (n � 14). The ADP/Oratio in the tumor samples was 1.2 � 0.3 (n � 7), whereas itwas 1.6 � 0.4 in the matched controls (n � 7), but thisdifference is not significant (Table 2).

ATP synthesis

Because the analysis of ATP synthesis can be usefullyperformed exclusively on mitochondria from fresh tissues,only the mitochondria from the seven tumors and theirmatched controls (n � 14), as used for the polarographicstudies, were used for ATP measurement. The mitochondrialATP synthesis, adjusted to the mitochondrial protein levelafter addition of the substrate, was 5.8 � 1.4 �mol/mg per10 min in oxyphilic tumors samples, compared with 12.1 �3.1 �mol/mg per 10 min in matched controls. This representsa low rate of mitochondrial ATP synthesis in comparisonwith the basal rate for normal thyroid tissue we had previ-ously measured (13.5 � 2.8) (unpublished data). The Wil-coxon test showed that the decrease observed in ATP syn-thesis was highly significant (P � 0.018, Table 2). There wasno difference in ATP levels between intact and permeabilized

mitochondria, attesting to the functionality of ATP/ADPtranslocase in exporting ATP from intact mitochondria.

MtDNA quantification

Twenty-two tumors and matched controls were exploredfor mtDNA quantification. DNA was preserved from rRNAcontamination by buffered RNase during extraction. Theratio of mtDNA to nDNA in oxyphilic tumor samples andmatched controls was determined by Southern blot analysis.Three hybridization bands were detected by the mitochon-drial 12S rRNA probe (7.5 and 1.7 kb) and the nuclear 18SrRNA probe (1.5 kb). The densitometric analysis of the 1.7-and 1.5-kb bands showed that the 12S rDNA/18S rDNA ratiowas 1.67 � 0.09 in oxyphilic thyroid tumors, compared with0.54 � 0.19 in controls. This increase in mtDNA content fortumors was highly significant, using the Wilcoxon test (P �0.001). Fig. 1 shows the 1.7- and 1.5-kb bands of severaloxyphilic tumors samples and matched controls.

Deletions in mtDNA were explored by long PCR analysisusing a standard procedure (16). The common mtDNA4977deletion was found in two of the tumors as well as in thematched controls, with an identical level of heteroplasmy.All the other samples were free from this deletion.

Mitochondrial and nuclear gene expression

Twenty-two tumors and matched controls were exploredfor mtDNA quantification. The expression of ND2 and ND5mitochondrial genes was 12 times higher in oxyphilic thyroidtumor samples than in controls. When adjusted to themtDNA/nDNA ratio, the relative mitochondrial transcriptratio was 3.8 times higher in the tumor samples than incontrols.

For the nuclear gene, we observed a 2-fold increase ofUCP2 in oxyphilic tumors samples, compared with controls.Fig. 2 shows the different patterns of mitochondrial andnuclear gene expression in the tumor samples and controls.Table 3 summarizes the histology and the gene expressionpattern of the different samples. Table 2 sums up the statis-tical analysis of the results.

Discussion

Several authors have suggested that defective energy-producing mechanisms of oxyphilic cells may be responsiblefor mitochondrial proliferation (11, 17). This hypothesisstems from the observation that the active metabolism of

TABLE 2. Statistical analysis of mitochondrial (ND2, ND5) and nuclear (UCP2) gene expression and ATP synthesis in oxyphilic thyroidtumors, compared with matched controls

Control tissuea

(n � 22)Oxyphilic tumorsa

(n � 22) Wilcoxon test

ND2b 658 � 290 6799 � 3215 P � 0.001ND5b 612 � 302 7354 � 3208 P � 0.001UCP2b 272 � 119 503 � 191 P � 0.001

n � 7 n � 7ADP/O ratio 1.6 � 0.4 1.2 � 0.3 NSATP (�mol/mg protein per 10 min) 12.1 � 3.1 5.8 � 1.4 P � 0.018

a Mean � SD.b Copy number/�-ACTIN copy number.ADP/O ratio, Rate of oxydative phosphorylation; NS, not significant.

4922 J Clin Endocrinol Metab, October 2001, 86(10):4920–4925 Savagner et al. • ATP Synthesis in Oxyphilic Thyroid Tumors

oncocytic cells, with their high levels of oxidative enzymes,does not correspond to high thyroid cell function (18). How-ever, in these histochemical studies, the respiratory enzymeswere functional, and a protein uncoupling the oxidativephosphorylation process (UCP1) was not present in oxyph-ilic tumors (11).

MtDNA alterations are associated with several mitochon-drial degenerative diseases (19). Large mtDNA deletions,such as the most common mtDNA4977 deletion, result in adecline of the oxidative phosphorylation capacity and accu-mulate progressively in aging normal tissues (20). An in-creased frequency of the mtDNA4977 deletion in oncocytictumors has been described (18). We were unable to identify

any mtDNA deletions that might provide a replicative advan-tage over wild-type mtDNA. The “common” mtDNA4977 de-letion was found in 2 of the 22 oxyphilic tumors investigated,but in each case the same deletion was also detected in thecorresponding controls and corresponded to two elderly pa-tients (63 and 75 yr). Thus, the mtDNA4977 deletion might beassociated with cellular aging rather than to the developmentof an oncocytic phenotype, as previously suggested (21, 22). Theincrease in mtDNA content observed in 22 thyroid tumors(3.10 � 0.29) was 25% lower than that indicated by other authors(4.31 � 1.09) (21).

The analysis of mitochondrial gene expression showedthat tumors had a 12-fold increase in ND2 and ND5 tran-scripts, compared with control tissue. The expression of thesetwo mitochondrial genes has already been associated withabnormal mitochondrial biogenesis in oncocytic tumors (12).The gene expression ratio was adjusted to the mtDNA/nDNA ratio calculated by Southern blot analysis to obtain amore accurate estimate of the real increase of mitochondrialgene expression. The mtRNA/mtDNA ratio thus deter-mined for the oxyphilic thyroid tumors may be comparedwith that given for oxyphilic tumors in other tissues. Wefound an mtRNA/mtDNA ratio of about 4:1 in oxyphilicthyroid tumors, compared with controls, whereas this ratiowas 1:1 in the case of oxyphilic salivary gland tumors and 1:5in the case of oxyphilic renal tumors (12). In the study of acell line derived from a thyroid oncocytoma, we found anmtRNA/mtDNA ratio as high as 2:1, compared with a con-trol thyroid cell line (23). These large differences in themtRNA/mtDNA ratio suggest tissue-specific regulation ofmitochondrial transcription and replication. It might there-fore be relevant to investigate the nuclear factors involved inthis regulation in various tissues.

Polarographic analysis produced no evidence of respi-ratory chain defects in oxyphilic thyroid tumors, com-pared with control tissue. The respiratory chain ratios inmitochondria isolated from seven oxyphilic tumors wereconsistent with the indices published for mitochondria inthe normal thyroid (24). However, the ADP/O ratio wasonly 75% of the normal value. The oxidative phosphory-lation coupling defect revealed by polarography might berelated to the 2-fold increase in UCP2 expression observedin oxyphilic tumors, compared with controls. After veri-fying that UCP1 was not expressed in oxyphilic thyroidtumors (data not shown), we investigated the expressionof UCP2, the role of which has been established in theuncoupling process (25).

FIG. 1. Quantitation of mtDNA andnDNA for several oxyphilic thyroid tu-mors and matched controls by Southernblot analysis. Five micrograms of totalDNA digested with XbaI was hybrid-ized first with an 18S rRNA nuclearprobe and then rinsed and hybridizedwith a 12S rRNA mtDNA probe. Theintensities of the 1.5-kb (18S) and1.7-kb (12S) bands were quantified by aRadioImager (Cyclone, Packard, Down-ers Grove, IL). C, Control tissue; O, oxy-philic thyroid tumor.

FIG. 2. Quantification of mitochondrial and nuclear gene transcriptsfrom 22 oxyphilic thyroid tumors and matched controls. 2a, Mean (�SD) of the ratio of the mitochondrial ND2 and ND5 cDNA copy num-bers vs. the �-ACTIN cDNA copy number. 2b, Mean (� SD) of the ratioof the nuclear UCP2 cDNA copy number vs. the �-ACTIN cDNA copynumber. The cDNA copies were determined by real-time quantitativePCR analysis (Lightcycler, Roche, Basel, Switzerland) after reversetranscription of the total RNA of each tumor and control sample.(Table 1 shows the PCR primers used.)

Savagner et al. • ATP Synthesis in Oxyphilic Thyroid Tumors J Clin Endocrinol Metab, October 2001, 86(10):4920–4925 4923

In our study of seven fresh oxyphilic tumors, the increasedexpression of UCP2 was probably responsible for the cou-pling defect reflected by a significant decrease in ATP syn-thesis. Mitochondrial proliferation could therefore be anadaptive response to a primary nuclear abnormality, namelythe overexpression of UCP2. However, the overexpression ofUCP2 might itself be a response to the proliferation of mi-tochondria compensating for the decreased mitochondrialATP synthesis. In the latest case, the proliferation of mito-chondria leads to overproduction of reactive oxygen species,which could be counteracted by an increase in UCP2 expres-sion (26).

Another line of reasoning suggests that the metabolism inoxyphilic tumors has switched to a glycolytic status (12). Thedecrease in mitochondrial ATP synthesis we noticed couldlead to a shift toward anaerobic metabolism. Because thedefect is measured in oxyphilic adenoma as well as in car-cinoma, we suggest that the metabolism switch is an earlyevent in the oxyphilic thyroid tumor progression. Thus, theoxyphilic cell might be early resistant to hypoxia, whichcould explain the aggressive clinical behavior of these tumors(22).

In conclusion, the defective ATP synthesis we observed inseven oxyphilic thyroid tumors might explain the mitochon-drial proliferation found in the tumor cells. Because the ex-pression of UCP2 was higher in all the 22 oxyphilic thyroidtumors, compared with controls, we suggest that the oxida-tive phosphorylation coupling defect we detected may beassociated with mitochondrial proliferation in oxyphilic thy-roid tumors. It would therefore be of interest to further in-vestigate the factors involved in the transcription and rep-lication of mitochondrial DNA in oxyphilic tumors of thethyroid gland.

Acknowledgments

We are grateful to C. Savagner for the statistical analysis and to K.Malkani for critical reading of the manuscript. We thank Anne, Domin-ique, and Florence for continuous support during the study.

Received January 26, 2001. Accepted June 5, 2001.Address all correspondence and requests for reprints to: F. Sav-

agner, Inserm EMI-U 00-18, Laboratoire de Biochimie et BiologieMoleculaire, Chu, 4 rue Larrey, F-49033 Angers cedex 01, France.E-mail: Frsavagner@chu-angers.fr.

This work was supported by grants from l’Association pour la Re-cherche sur le Cancer.

References

1. Hedinger C, Williams ED, Sobin LH 1989 The WHO histological classificationof thyroid tumors: a commentary on the ed. 2. Cancer 63:908–911

2. Katoh R, Harach HR, Williams ED 1998 Solitary, multiple, and familial oxy-phil tumours of the thyroid gland. J Pathol 186:292–299

3. Ricquier D, Bouillaud F 2000 The uncoupling protein homologues: UCP1,UCP2, UCP3, StUCP and AtUCP. Biochem J 345:161–179

4. Tallini G 1998 Oncocytic tumours. Virchows Arch 433:5–125. Bronner MP, Clevenger CV, Edmonds PR, Lowell DM, McFarland MM,

LiVolsi VA 1988 Flow cytometric analysis of DNA content in Hurthle celladenomas and carcinomas of the thyroid. Am J Clin Pathol 89:764–769

6. Bouillaud F, Ricquier D, Thibault J, Weissenbach J 1985 Molecular approachto thermogenesis in brown adipose tissue: cDNA cloning of the mitochondrialuncoupling protein. Proc Natl Acad Sci USA 82:445–448

7. DiMauro S, Bonilla E, Zeviani M, Nakagawa M, DeVivo DC 1985 Mito-chondrial myopathies. Ann Neurol 17:521–538

8. Shoffner JM, Wallace DC 1990 Oxidative phosphorylation diseases. Disordersof two genomes. Adv Hum Genet 19:267–330

9. Heddi A, Lestienne P, Wallace DC, Stepien G 1993 Mitochondrial DNAexpression in mitochondrial myopathies and coordinated expression of nu-clear genes involved in ATP production. J Biol Chem 268:12156–12163

10. Flier J, Lowell B 1997 Obesity research springs a proton leak. Nat Genet15:223–224

11. Ebner D, Rodel G, Pavenstaedt I, Haferkamp O 1991 Functional and mo-lecular analysis of mitochondria in thyroid oxyphil tumour. Virchows Arch BCell Pathol Incl Mol Pathol 60:139–144

12. Heddi A, Faure-Vigny H, Wallace DC, Stepien G 1996 Coordinate expressionof nuclear and mitochondrial genes involved in energy production in carci-noma and oxyphil tumour. Biochim Biophys Acta 1316:203–209

13. Chretien D, Bourgeron T, Rotig A, Munnich A, Rustin P 1990 The measure-

TABLE 3. Case reports

Case HistologyND2# ND5# UCP2#

C O C O C O

1 OTA 752 7,890 812 6,850 355 4692 OFA 267 5,749 320 7,760 123 1983 OFC 402 2,592 130 1,390 420 9724 OFA 1,140 14,587 1,250 11,760 201 4955 OFA 760 8,403 548 11,540 341 5076 OFC 804 8,702 647 12,501 201 3577 OFA 589 3,015 950 5,470 258 3988 OTA 549 4,057 450 5,690 507 9809 OFA 705 9,587 521 12,589 204 480

10 OFA 941 5,783 203 2,547 438 50811 OTA 875 2,378 875 7,646 321 54212 OTA 1,236 12,700 694 11,345 365 60313 OTA 865 5,640 549 5,578 532 68514 OTA 1,130 6,389 253 3,543 130 60315 OFA 456 8,653 159 3,211 147 36516 OFA 442 7,833 377 6,489 230 52117 OFC 583 6,421 590 7,345 184 47518 OFA 345 10,054 1,050 8,897 163 29519 OFA 321 3,890 858 6,789 208 26520 OFA 238 3,420 665 5,780 201 38521 OFA 643 7,804 947 8,085 220 43022 OFA 436 4,042 635 8,988 251 541

#, Copy number/�-ACTIN copy number; OFA, oxyphilic follicular adenoma; OFC, oxyphilic follicular carcinoma; OTA, oxyphilic follicularadenoma (trabecular pattern); C, control surrounding tissue; O, oxyphilic thyroid tumor.

4924 J Clin Endocrinol Metab, October 2001, 86(10):4920–4925 Savagner et al. • ATP Synthesis in Oxyphilic Thyroid Tumors

ment of the rotenone-sensitive NADH cytochrome c reductase activity inmitochondria isolated from minute amount of human skeletal muscle. Bio-chem Biophys Res Commun 173:26–33

14. Lemasters JJ, Hackenbrock CR 1976 Continuous measurement and rapidkinetics of ATP synthesis in rat liver mitochondria, mitoplasts and innermembrane vesicles determined by firefly-luciferase luminescence. Eur J Bio-chem 67:1–10

15. Tallini G, Carcangiu ML, Rosai J 1992 Oncocytic neoplasms of the thyroidgland. Acta Pathol Jpn 42:305–315

16. Reynier P, Chretien MF, Savagner F, et al. 1998 Long PCR analysis of humangamete mtDNA suggests defective mitochondrial maintenance in spermato-zoa and supports the bottleneck theory for oocytes. Biochem Biophys ResCommun 254:373–377

17. Heddi A, Stepien G, Benke PJ, Wallace DC 1999 Coordinate induction ofenergy gene expression in tissues of mitochondrial disease patients. J BiolChem 274:22968–22976

18. Welter C, Kovacs G, Seitz G, Blin N 1989 Alteration of mitochondrial DNAin human oxyphil tumours. Genes Chromosomes Cancer 1:79–82

19. Wallace DC 1992 Diseases of the mitochondrial DNA. Annu Rev Biochem61:1175–1212

20. Cortopassi G, Shibata D, Soong N, Arnheim N 1992 A pattern of accumu-lation of a somatic deletion of mitochondrial DNA in aging human tissues. ProcNatl Acad Sci USA 89:7370–7374

21. Tallini G, Ladanyi M, Rosai J, Jhanwar SC 1994 Analysis of nuclear andmitochondrial DNA alterations in thyroid and renal oncocytic tumors. Cyto-genet Cell Genet 66:253–259

22. Maximo V, Sobrinho-Simoes M 2000 Hurthle cell tumours of the thyroid. Areview with emphasis on mitochondrial abnormalities with clinical relevance.Virchows Arch 437:107–115

23. Savagner F, Chevrollier A, Loiseau D, et al. 2001 Mitochondrial activity inXTC.UC1 cells derived from thyroid oncocytoma. Thyroid 11:327–333

24. Lamy FM, Rodesch FR, Dumont JE 1967 Action of thyrotropin on thyroidenergetic metabolism. VI. Regulation of mitochondrial respiration. Exp CellRes 46:518–532

25. Ricquier D, Bouillaud F 2000 Mitochondrial uncoupling proteins: from mi-tochondria to the regulation of energy balance. J Physiol 529:3–10

26. Arsenijevic D, Onuma H, Pecqueur C, et al. 2000 Disruption of the uncouplingprotein-2 gene in mice reveals a role in immunity and oxygen species pro-duction. Nat Genet 26:435–439

Savagner et al. • ATP Synthesis in Oxyphilic Thyroid Tumors J Clin Endocrinol Metab, October 2001, 86(10):4920–4925 4925