Pheochromocytomas in von Hippel-Lindau Syndromeand Multiple Endocrine Neoplasia Type 2 DisplayDistinct Biochemical and Clinical Phenotypes

GRAEME EISENHOFER, MCCLELLAN M. WALTHER, THANH-TRUC HUYNH,SHENG-TING LI, STEFAN R. BORNSTEIN, ALEXANDER VORTMEYER,MASSIMO MANNELLI, DAVID S. GOLDSTEIN, W. MARSTON LINEHAN,JACQUES W. M. LENDERS, AND KAREL PACAK

Clinical Neurocardiology Section and Surgical Neurology Branch, National Institute of NeurologicalDisorders and Stroke (G.E., T.-T.H., S.-T.L., A.V., D.S.G.); Urologic Oncology Branch, National CancerInstitute (M.M.W., W.M.L.); and Pediatric and Reproductive Endocrinology Branch, National Instituteof Child Health and Human Development (S.R.B., K.P.), National Institutes of Health, Bethesda,Maryland 20892; Department of Clinical Pathophysiology, University of Florence (M.M.), Florence,Italy; and Department of Internal Medicine, St. Radboud University Hospital (J.W.M.L.), Nijmegen,The Netherlands

ABSTRACTThis study examined the mechanisms linking different biochemical

and clinical phenotypes of pheochromocytoma in multiple endocrineneoplasia type 2 (MEN 2) and von Hippel-Lindau (VHL) syndrome tounderlying differences in the expression of tyrosine hydroxylase (TH),the rate-limiting enzyme in catecholamine synthesis, and of phe-nylethanolamine N-methyltransferase (PNMT), the enzyme that con-verts norepinephrine to epinephrine. Signs and symptoms of pheo-chromocytoma, plasma catecholamines and metanephrines, andtumor cell neurochemistry and expression of TH and PNMT wereexamined in 19 MEN 2 patients and 30 VHL patients with adrenalpheochromocytomas. MEN 2 patients were more symptomatic andhad a higher incidence of hypertension (mainly paroxysmal) andhigher plasma concentrations of metanephrines, but paradoxically

lower total plasma concentrations of catecholamines, than VHL pa-tients. MEN 2 patients all had elevated plasma concentrations of theepinephrine metabolite, metanephrine, whereas VHL patientsshowed specific increases in the norepinephrine metabolite, normeta-nephrine. The above differences in clinical presentation were largelyexplained by lower total tissue contents of catecholamines and ex-pression of TH and negligible stores of epinephrine and expression ofPNMT in pheochromocytomas from VHL than from MEN 2 patients.Thus, mutation-dependent differences in the expression of genes con-trolling catecholamine synthesis represent molecular mechanismslinking the underlying mutation to differences in clinical presentationof pheochromocytoma in patients with MEN 2 and the VHL syndrome.(J Clin Endocrinol Metab 86: 1999–2008, 2001)

PHEOCHROMOCYTOMAS are tumors of chromaffincells, typically arising in the adrenal gland and char-

acterized by excess production of catecholamines. Most tu-mors secrete predominantly norepinephrine, many produceboth norepinephrine and epinephrine, and more rarely oth-ers secrete predominantly epinephrine (1, 2). These differ-ences in norepinephrine and epinephrine secretion can ex-plain differences in presenting symptoms (2–4).

In some pheochromocytomas catecholamine secretion ap-pears to be continuous, whereas in others, particularly epi-nephrine-secreting tumors, secretion is episodic (2, 5). Suchdifferences may account for why some patients with pheo-chromocytoma present with sustained hypertension,whereas others present with paroxysmal hypertension andattacks of sweating, tachycardia, or anxiety. In some patientswith pheochromocytoma, particularly those in whom thetumor is discovered during periodic screening for hereditary

pheochromocytoma or as an incidentaloma during imagingstudies for other medical conditions, the tumor may notproduce signs or symptoms (6–10). In this setting, plasmaand urinary catecholamines often are normal, indicating littlesecretion of the amines by the tumor.

We recently reported that measurements of plasma con-centrations of normetanephrine and metanephrine, the O-methylated metabolites of norepinephrine and epinephrine,provide a particularly sensitive test for detecting pheochro-mocytomas in patients with von Hippel-Lindau (VHL) syn-drome or multiple endocrine neoplasia type 2 (MEN 2) (11).These multisystem neoplastic disorders are inherited in anautosomal dominant fashion and account for most currentlyidentified hereditary pheochromocytomas. The continuousnature of production of metanephrines within tumor tissue,production that is independent of catecholamine secretion,revealed that pheochromocytomas from patients with MEN2 produce metanephrine, whereas those from VHL patientsproduce almost exclusively normetanephrine (11). Thesefindings suggest that pheochromocytomas in the VHL syn-drome are characterized by a noradrenergic biochemicalphenotype, whereas those in patients with MEN 2 are char-acterized by an adrenergic phenotype.

Received August 25, 2000. Revision received January 8, 2001. Ac-cepted January 14, 2001.

Address all correspondence and requests for reprints to: Dr. GraemeEisenhofer, Building 10, Room 6N252, National Institutes of Health, 10Center Drive, MSC 1620, Bethesda, Maryland 20892-1620. E-mail:ge@box-g.nih.gov.

0021-972X/01/$03.00/0 Vol. 86, No. 5The Journal of Clinical Endocrinology & Metabolism Printed in U.S.A.Copyright © 2001 by The Endocrine Society

1999

Although previous studies indicated that many patientswith MEN 2 have epinephrine-secreting pheochromocyto-mas (9, 12–14), the basis for this is not established, and thesestudies did not compare tumor phenotypes in patients withMEN 2 and VHL syndrome. This study examined the hy-pothesis that pheochromocytomas in MEN 2 and VHL pa-tients exhibit specific noradrenergic vs. adrenergic biochem-ical phenotypes that reflect mutation-dependent differentialexpression of genes regulating catecholamine synthesis.More specifically this study aimed to establish whether dif-ferences in the expression of tyrosine hydroxylase (TH), therate-limiting enzyme in catecholamine synthesis, and of phe-nylethanolamine N-methyltransferase (PNMT), the enzymethat converts norepinephrine to epinephrine, might accountfor different biochemical phenotypes and clinical presenta-tions of pheochromocytomas in patients with MEN 2 andVHL syndrome.

Subjects and MethodsSubjects

The patient database for this report includes 49 patients with histo-logically proven adrenal pheochromocytomas, 30 (12 women and 18men) associated with VHL syndrome and 19 (9 women and 10 men) withMEN 2 (18 with MEN 2a and 1 with MEN 2b). Three VHL patients and1 MEN 2 patient had pheochromocytomas removed on 2 separate oc-casions, between 1–4 yr apart, giving a total of 53 cases of adrenalpheochromocytoma. At the time of tumor resection, VHL patients hada mean (6sd) age of 29 6 14 yr (range, 8–63 yr) and MEN 2 patients wereaged 36 6 10 yr (range, 22–53 yr). The diagnosis of VHL syndrome orMEN 2 was confirmed by identification of germline mutations of theVHL tumor suppressor gene or the RET protooncogene in all patients.The absence or presence of symptoms attributable to pheochromocy-toma (e.g. headache, diaphoresis, and palpitations) or of hypertension,and whether hypertension was sustained or paroxysmal, were assessedby review of patient records. Studies were approved by the appropriateinstitutional review boards, and all patients gave informed consent toparticipate.

Collection of blood and tissue samples

Blood samples were obtained from all patients using an indwellingiv catheter inserted into a forearm vein, with patients supine for at least20 min before blood collection. Samples of blood were transferred intotubes containing heparin as anticoagulant and immediately placed onice until centrifuged (4 C) to separate the plasma. Plasma samples werestored at 280 C until assayed.

Samples of tumor tissue were obtained at surgery from 18 VHLpatients and 12 MEN 2 patients within 1 h of surgical removal. Tumorswere placed on ice immediately after removal, extraneous tissue wasremoved, dimensions of tumors were recorded, and small samples (50–400 mg) were dissected away from surrounding tissue, placed on dry iceor in liquid nitrogen, and then stored at 280 C.

Plasma and tissue catecholamines and metanephrines

Plasma and tissue concentrations of catecholamines (norepinephrine,epinephrine and dopamine) were quantified by liquid chromatographywith electrochemical detection. Samples of tissue were weighed andhomogenized in at least 5 vol 0.4 mol/L perchloric acid containing 0.5mmol/L ethylenediamine tetraacetate. Homogenized samples were cen-trifuged (1500 3 g for 15 min at 4 C), and supernatants collected andstored at 280 C until assayed. Concentrations of catecholamines weredetermined after extraction from plasma or perchloric acid tissue su-pernatants using alumina adsorption as described previously (15).

Plasma concentrations of metanephrines (normetanephrine andmetanephrine) were determined using a different liquid chromatogra-phy procedure after extraction onto solid phase ion exchange columns(16).

Intraassay coefficients of variation were 1.9% for norepinephrine,3.0% for epinephrine, 4.2% for normetanephrine, and 3.3% for meta-nephrine. Interassay coefficients of variation were 3.2% for norepineph-rine, 9.9% for epinephrine, 7.1% for normetanephrine, and 5.1% formetanephrine.

Tissue TH activity

The activity of TH in samples of pheochromocytoma tumor tissue wasdetermined from measurements of the formation dihydroxyphenylala-nine from tyrosine according to a previously described assay (17). Inbrief, an appropriately diluted sample of each tissue preparation wasincubated with 0.1 mol/L acetate buffer (pH 6.0), 200 mmol/L l-tyrosine,1.25 mmol/L m-hydroxybenzylhydrazine, 1 mmol/L d,l-6-methyltetra-hydropterine dihydrochloride, 19.5 3 103 U/mL catalase, 3.8 U/mLdihydropteridine reductase, 1 mmol/L NADPH, and 0.4 mol/L glycerolto a total volume of 200 mL. The mixture was incubated at 37 C for 10min, the incubation was terminated by adding 100 mL 0.4 mol/L per-chloric acid, and dihydroxyphenylanine was quantified by liquid chro-matography with electrochemical detection. TH activity (picomoles permin/mg wet wt tissue) was calculated using the formula: TH activity 5(DOPAsample 2 DOPAblank)/(weight of tissue 3 time of incubation).

Western blot analysis of TH and PNMT

Cytosolic proteins were prepared according to the procedure of An-drews and Faller (18). Samples of pheochromocytoma tissue (;15 mg)were homogenized using a Dounce homogenizer (Kontes Co., Vineland,NJ) in 0.5 mL 10 mmol/L HEPES-KOH buffer (pH 7.9) containing 1.5mmol/L MgCl2, 10 mmol/L KCl, 0.5 mmol/L dithiothreitol, 0.2mmol/L phenylmethylsulfonylfluoride, and protease inhibitors. Cyto-solic fractions were separated from pelleted debris by centrifugation at17,000 3 g for 10 s (4 C).

Cytosolic proteins (20 mg) were then electrophoresed on a 12% SDS-polyacrylamide gel for PNMT and on a 10% SDS-polyacrylamide gel forTH and transferred onto polyvinylidene fluoride membranes (MilliporeCorp., Bedford, MA) using a transblot apparatus (Bio-Rad Laboratories,Inc., Hercules, CA). After transfer, polyvinylidene fluoride membraneswere incubated in blocking buffer [50 mmol/L Tris (pH 7.4), 0.9% NaCl,0.05% Tween-20, and 10% dry milk] overnight at 4 C. Membranes werethen washed three times for 10 min each time with Tris-buffered saline[50 mmol/L Tris (pH 7.4) and 0.9% NaCl] containing 0.05% Tween-20,and then incubated with either rabbit anti-PNMT polyclonal antibody(1:1000 dilution; Chemicon, Temecula, CA) or mouse anti-TH mono-clonal antibody (1:4000 dilution; Calbiochem, San Diego, CA) for 1 h atroom temperature.

Membranes were washed again three times for 10 min each time withTris-buffered saline containing 0.05% Tween-20 and then incubated for1 h with horseradish peroxidase-conjugated antirabbit IgG or antimouseIgG at a 1:20,000 dilution for both antibodies. Membranes were againwashed three times for 10 min each time with Tris-buffered salinecontaining 0.05% Tween-20. PNMT and TH bands were visualized usingthe enhanced chemiluminescence method.

Quantitative PCR analysis of TH and PNMT

Ribonucleic acid (RNA) was extracted from pheochromocytoma tis-sue using TRIzol reagent (Life Technologies, Inc., Gaithersburg, MD).Traces of DNA were removed by digestion with deoxyribonuclease-freeribonuclease (Gene Hunter, Nashville, TN). Total RNA (1 mg) was re-versibly transcribed to complementary DNA (cDNA) using randomhexamers together with the Superscript Preamplification System forFirst Strand cDNA Synthesis (Life Technologies, Inc.). Real Time Quan-titative PCR (TaqMan PCR), using a 7700 Sequence Detector (Perkin-Elmer Corp./PE Applied Biosystems, Foster City, CA), was used forquantification of PNMT or TH messenger RNA (mRNA) as describedpreviously (19). The amounts of PNMT and TH mRNA were determinedby amplification of the cDNA target using the following primers andTaqMan probes designed from the human PNMT or TH gene sequencesby the Primer Express program from Perkin-Elmer Corp./PE AppliedBiosystems: PNMT forward primer, 59-GCA GCC ACT TTG AGG ACATCA-39; PNMT reverse primer, 59-GGC TGT ACA TGC TCC AGT TGAA-39; PNMT TaqMan probe, 59 (FAM)-CAG ATT TCC TGG AGG TCA

2000 EISENHOFER ET AL. JCE & M • 2001Vol. 86 • No. 5

ACC GCC A-(TAMRA) 39; TH forward primer, 59-CGG ATG AGG AAATTG AGA AGC T-39; TH reverse primer, 59-TCT GCT TAC ACA GCCCGA ACT-59; and TH TaqMan probe, 59 (FAM)-CCA CGC TGT CATGGT TCA CGG TG-(TAMRA) 39.

To normalize quantification of PNMT or TH mRNA for differencesin the amount of total RNA added to each cDNA reaction, 18S ribosomalRNA served as a housekeeping gene, which was detected using theTaqMan Ribosomal RNA Control Reagents (Perkin-Elmer Corp./PEApplied Biosystems). To minimize random errors, PCR amplification ofPNMT or TH genes and 18S ribosomal RNA was carried out in the sametube. Reaction mixtures contained 5 mL cDNA product as template, 1 3TaqMan Universal PCR Master Mix, 3 mmol/L for each PNMT or THforward and reverse primer, 2 mmol/L for the PNMT or TH TaqManprobe, 10 mmol/L for each 18S forward and reverse primer, 40 mmol/Lfor the 18S TaqMan probe, and water to a final volume of 50 mL. Thefollowing temperature parameters were cycled 50 times: 15 s at 95 C and1 min at 60 C. Input RNA amounts were calculated manually using theComparative CT method for both target genes and 18S. The amount ofPNMT or TH mRNA was normalized by division by the amount of 18SRNA in each sample.

Electron microscopy

Pheochromocytoma tissue was fixed for 3 h in 2% formaldehyde and2% glutaraldehyde in 0.1 mol/L phosphate buffer, pH 7.3. Tissue sliceswere postfixed for 90 min (2% OsO4 in 0.1 mol/L cacodylate buffer, pH7.3), dehydrated in ethanol, and embedded in epoxy resin. Ultrathinsections were stained with uranyl acetate and lead citrate and examinedand photographed at 80 kV in a Phillips CM10 electron microscope(Phillips Electronic Instruments, Mahway, NJ).

Statistics

The distributions of plasma and tumor tissue concentrations of cat-echolamines and plasma concentrations of metanephrines in patientswith pheochromocytoma were highly skewed. Normal distributionswere obtained after logarithmic transformation of the data. Mean valuesfor these variables are therefore provided as geometric means. Corre-sponding ses were established from the normalized data. All otherresults for normally distributed data are expressed as the arithmeticmean 6 sem. Where data showed nonnormal distributions, statisticaltests of significance were carried out on normalized data. These testsincluded paired t tests and ANOVAs with post-hoc tests carried out usingthe Scheffe F test. x2 analysis was used to examine differences in pre-senting signs and symptoms. Differences among relationships betweenplasma concentrations of catecholamines or between tumor size or cat-echolamine content and the presence or absence of symptoms or hy-pertension were examined by multiple linear regression analysis.

ResultsPheochromocytoma tissue catecholamines

Pheochromocytomas from VHL patients displayed a dis-tinctly and consistently noradrenergic phenotype, with nor-epinephrine concentrations representing 98.0 6 0.4% andepinephrine concentrations only 1.5 6 0.4% of the total cat-echolamine content (Fig. 1A). In contrast, epinephrine ac-counted for 47.5 6 6.3% and norepinephrine for 52.3 6 6.3%of the total catecholamine content of pheochromocytomatumor tissue in patients with MEN 2. Dopamine was a minorcomponent, amounting to less than 0.5% of the total cate-cholamine content of tumors from VHL and MEN 2 patients.

Concentrations of catecholamines (the sum of dopamine,norepinephrine, and epinephrine) in pheochromocytoma tis-sue varied widely from 2–170 mmol/g, but averaged 4.8-foldhigher (P , 0.001) in tumors from MEN 2 patients than inthose from VHL patients (69.2 vs. 14.4 mmol/g). This differ-ence was largely due to concentrations of epinephrine thatwere 235-fold higher (P , 0.001) in tumors from MEN 2

patients than in those from VHL patients (Fig. 1A). However,patients with MEN 2 also had 2.3-fold higher tumor tissueconcentrations of norepinephrine (P 5 0.019) and 3.7-fold

FIG. 1. Concentrations of norepinephrine (o) and epinephrine (f) inpheochromocytoma tumor tissue (A) or plasma (B) and concentrationsof normetanephrine (o) and metanephrine (f) in plasma (C) of VHLpatients (left) compared with MEN 2 patients (right). Results are thegeometric mean 6 SEM. *, Significantly (P , 0.02) different value inMEN 2 patients than in VHL patients.

PHEOCHROMOCYTOMA PHENOTYPES 2001

higher concentrations of dopamine (P , 0.001) than VHLpatients.

Plasma catecholamines and metanephrines

Plasma concentrations of norepinephrine were 2-foldhigher (P 5 0.009) in VHL patients than in MEN 2 patientswith pheochromocytoma (Fig. 1B). In contrast, plasmaconcentrations of epinephrine were 5-fold higher (P ,0.001) in MEN 2 patients than in VHL patients withpheochromocytoma.

Unlike the pattern for norepinephrine, plasma concentra-tions of normetanephrine did not differ among VHL andMEN 2 patients with pheochromocytoma (Fig. 1C). Plasmaconcentrations of metanephrine, however, were 16-foldhigher (P , 0.001) in MEN 2 patients than in VHL patientswith pheochromocytoma.

The predominant features distinguishing the biochemicaldiagnostic presentation of the two hereditary pheochromo-cytoma syndromes were higher plasma concentrations ofmetanephrine and epinephrine in MEN 2 patients than inVHL patients (Fig. 2). However, half of all MEN 2 patientswith pheochromocytoma had normal plasma epinephrineconcentrations, whereas all had elevated plasma concentra-tions of metanephrine. Very few patients with VHL diseasehad elevations of either plasma epinephrine (3%) or meta-nephrine (9%). In the three VHL patients who had elevatedplasma concentrations of metanephrine, the increases abovethe upper reference limit of normal were slight (,20%). Thus,whereas there was considerable overlap in plasma concen-trations of epinephrine among VHL and MEN 2 patients,there was no overlap in plasma concentrations ofmetanephrine.

Expression of TH

Quantitative TaqMan PCR revealed that PNMT mRNAwas expressed in pheochromocytoma tumor tissue fromVHL patients at 22% the level of expression in tumor tissuefrom patients with MEN 2 (P , 0.001; Fig. 3A). Similarly,Western blot analysis showed lower levels of expression ofTH protein in pheochromocytoma tissue from VHL patientsthan from MEN 2 patients (Fig. 3B). Moreover, levels of THenzyme activity in tumor tissue from VHL patients were 19%(P , 0.002) those observed in tissue from patients with MEN2 (Fig. 3C). Levels of TH mRNA correlated positively withTH enzyme activity (r 5 0.62; P 5 0.011) and total catechol-amine contents of tumors (r 5 0.69; P 5 0.006).

Expression of PNMT

Quantitative TaqMan PCR revealed that PNMT mRNAwas expressed in pheochromocytoma tumor tissue fromVHL patients at less than 2% (P , 0.001) the level of expres-sion in tumor tissue from patients with MEN-2 (Fig. 4A).Similarly, Western blot analysis showed consistent expres-sion of PNMT protein in pheochromocytoma tissue fromMEN-2 patients and a general lack of expression in VHLpatients (Fig. 4B).

Tumor cell morphology

Electron microscopic analysis revealed distinct ultrastruc-tural differences between pheochromocytoma tumor cellsfrom VHL and MEN 2 patients (Fig. 5). Chromaffin tumorcells from patients with MEN 2 shared many of the charac-teristics of normal adrenal medullary chromaffin cells,whereas tumor cells from VHL patients did not. The cyto-plasm of MEN 2 tumor cells was filled with two types ofsecretory granules in similar amounts: 1) epinephrine-containing large, round or elongated, medium density gran-ules with a particulate substructure; and 2) small norepi-nephrine-containing, electron-dense granules (Fig 5, A andC). In MEN 2 tumor cells the two types of secretory granuleswere evenly distributed throughout the cytoplasm, and insome cells there were enlarged round mitochondria. In con-trast, pheochromocytoma tumor cells from VHL patientscontained fewer granules than did MEN 2 tumor cells, andmost vesicles exhibited a dense core with a large lucent halotypical of norepinephrine-containing granules (Fig 5, B andD). Pheochromocytoma tumor cells from VHL patientsshowed an increased amount of rough endoplasmic reticu-lum, and the secretory granules were most frequently linedup for exocytosis along cell membranes.

Relationships of tumor size and catecholamine content toplasma catecholamines and metanephrines

Plasma concentrations of total metanephrines (combinedsum of plasma normetanephrine and metanephrine) andtotal catecholamines (combined sum of plasma norepineph-rine and epinephrine) were positively correlated (P , 0.001)with tumor size (Fig. 6). The relationships were stronger formetanephrines (r 5 0.87) than for catecholamines (r 5 0.69)and differed among VHL and MEN 2 patients. Relative totumor size, patients with MEN 2 had higher (P 5 0.010)

FIG. 2. Plasma concentrations of epinephrine compared with meta-nephrine in VHL patients and MEN 2 patients with pheochromocy-toma. Each point shows the concentration of metanephrine or epi-nephrine in an individual patient. The dashed horizontal lines showthe upper reference limits of normal for plasma concentrations ofepinephrine (0.45 pmol/mL) and metanephrine (0.31 pmol/mL).

2002 EISENHOFER ET AL. JCE & M • 2001Vol. 86 • No. 5

plasma concentrations of total metanephrines, but lower (P ,0.001) plasma concentrations of total catecholamines thandid VHL patients.

Plasma concentrations of total metanephrines and total

catecholamines were also positively correlated with the totalcatecholamine content of tumors, but relationships weremuch stronger for metanephrines (r 5 0.79; P , 0.001) thanfor catecholamines (r 5 0.48; P 5 0.027; Fig. 7). Relationshipsbetween total tumor catecholamine content and plasma totalmetanephrines were similar for VHL and MEN 2 patients,whereas relative to tumor catecholamine content, VHL pa-tients had higher (P 5 0.019) plasma concentrations of cat-echolamines than did MEN 2 patients.

Signs and symptoms of pheochromocytoma

Many of the patients with pheochromocytoma, particu-larly the VHL patients, were normotensive and asympto-matic (Table 1). Only 18% of VHL patients with pheochro-mocytoma presented with hypertension, which was usuallypersistent. In contrast, hypertension was present in 40% ofpatients with MEN 2 and was usually paroxysmal. Thus,paroxysmal hypertension was considerably more common(P 5 0.005) in MEN 2 patients than in VHL patients.

Symptoms attributable to pheochromocytoma were re-ported twice as frequently (P 5 0.033) in MEN 2 patients asin VHL patients (Table 1). Moreover, MEN 2 patients re-ported that these symptoms occurred as distinct attacks or

FIG. 3. Expression of TH mRNA (A), levels of TH protein by Westernblot (B), and levels of TH enzyme activity (C) in pheochromocytomatumor tissue from patients with VHL syndrome or MEN 2. Levels ofTH mRNA were determined by quantitative TaqMan PCR in pheo-chromocytomas from nine VHL and eight MEN 2 patients and areexpressed relative to the levels of 18S RNA. Expression of TH proteinby Western blot shows bands corresponding in molecular weight totwo isoforms of TH in representative samples of tumor tissue from sixpatients with MEN 2 compared with six VHL patients. Levels of THenzyme activity were determined in pheochromocytomas from nineVHL and eight MEN 2 patients.

FIG. 4. Expression of PNMT mRNA (A) and protein (B) in pheochro-mocytoma tumor tissue from patients with VHL syndrome or MEN 2.Levels of PNMT mRNA were determined by quantitative TaqManPCR in tumor tissue from 11 VHL and 6 MEN 2 patients and areexpressed relative to the levels of 18S RNA. Expression of PNMTprotein, determined by Western blot, shows bands corresponding inmolecular weight to PNMT in representative samples of tumor tissuefrom 6 patients with MEN 2 compared with lack of expression in 6VHL patients (2 isoforms of PNMT are apparent in several samples).

PHEOCHROMOCYTOMA PHENOTYPES 2003

episodes over three times more frequently (P 5 0.006) thanVHL patients. The most commonly reported symptoms wereheadache, diaphoresis, palpitations, and anxiety. Other lesscommon symptoms were tiredness or fatigue (n 5 8), diz-ziness or faintness (n 5 5), facial flushing (n 5 4), nausea withor without vomiting (n 5 3), constipation (n 5 3), and trem-ulousness (n 5 1). Apart from constipation, which was onlyreported in three VHL patients, none of these symptomsappeared to distinguish VHL from MEN 2 patients.

Relationships of signs and symptoms to tumor size andplasma catecholamines

Tumor size was a significant determinant of hypertension(P 5 0.005) and the presence of symptoms of pheochromo-cytoma (P 5 0.002) for the combined data from VHL andMEN 2 patients. However, the influence of tumor size on thepresence of hypertension and symptoms tended to be stron-ger in MEN 2 patients than in VHL patients, so that all MEN2 patients with tumors larger than 5 cm in average diameter

presented with hypertension and symptoms compared withonly 50% of VHL patients (Fig. 8A).

The concentration of total catecholamines (sum of norepi-nephrine and epinephrine) in plasma was also a significantdeterminant of hypertension (P 5 0.007) and symptoms (P 50.017) for the combined data from patients with VHL diseaseand MEN 2. However, whereas this represented a stronginfluence in patients with MEN 2, plasma concentrations ofcatecholamines had remarkably little influence on hyperten-sion or symptoms in VHL patients (Fig. 8B). Thus, relative toplasma concentrations of total catecholamines, patients withMEN 2 had a much higher frequency of hypertension (P 50.015) and symptoms of pheochromocytoma (P 5 0.010) thandid VHL patients.

Discussion

This study highlights and reveals the mechanisms respon-sible for different clinical manifestations of pheochromocy-toma in MEN 2 and VHL patients. Distinct mutation-depen-

FIG. 5. Electron micrographs of pheochromocytoma tumor cells at low (A and B) and high (C and D) magnifications from a patient with MEN2 (A and C) compared with a VHL patient (B and D). On the ultrastructural level at low magnification (bar, 0.5 mm), the distribution ofcatecholamine secretory granules is more sparse in VHL tumor cells (B) than in MEN 2 tumor cells (A). Storage granules in MEN 2 tumor cellsare evenly distributed throughout the cytoplasm, whereas in VHL tumor cells they are distributed lined up and in close proximity to the cellmembrane (see arrows). At high magnification (bar, 0.03 mm), MEN 2 tumor cells are characterized by the presence of both epinephrine (E)and norepinephrine (NE) secretory granules (C), whereas VHL tumor cells contain predominantly norepinephrine granules with the charac-teristic eccentrically situated, electron-dense core.

2004 EISENHOFER ET AL. JCE & M • 2001Vol. 86 • No. 5

dent biochemical phenotypes explain many of thedifferences in clinical presentation.

Pheochromocytomas in patients with MEN 2 expressPNMT and thus display an adrenergic biochemical pheno-type, whereas those in VHL patients express negligiblePNMT and therefore show a distinctly noradrenergic bio-chemical phenotype. Consequently, MEN 2 patients withpheochromocytoma all show elevations in plasma meta-nephrine, the metabolite of epinephrine, whereas VHL pa-tients typically show elevations only in normetanephrine, themetabolite of norepinephrine. Moreover, due to greater ex-pression of TH, and thus higher rates of catecholamine bio-synthesis, pheochromocytomas in MEN 2 patients containconsiderably larger amounts of catecholamines than those inVHL patients. Consequently, MEN 2 patients show largerelevations of plasma metanephrines than VHL patients andhave a higher frequency of hypertension and symptoms,

particularly of a paroxysmal nature. Paradoxically, however,basal plasma levels of total catecholamines are lower in MEN2 patients with pheochromocytoma than in VHL patients,indicating that pheochromocytomas from VHL and MEN 2

FIG. 6. Relationships between tumor size and plasma concentrationsof total metanephrines (A) and total catecholamines (B) in VHL pa-tients (E) compared with MEN 2 patients (F). Relationships for VHLpatients are shown by dashed regression lines, and those for MEN 2patients are shown by solid regression lines. Plasma concentrationsof total metanephrines represent the sum of plasma concentrations ofnormetanephrine and metanephrine. Plasma concentrations of totalcatecholamines represent the sum of plasma concentrations ofnorepinephrine and epinephrine. Multiple regression analysis re-vealed that relative to tumor size, patients with MEN 2 have higher(P 5 0.010) plasma concentrations of total metanephrines and lower(P , 0.001) plasma concentrations of total catecholamines than VHLpatients.

FIG. 7. Relationships between tumor catecholamine content andplasma concentrations of total metanephrines (A) and total cat-echolamines (B) in VHL patients (E) compared with MEN 2 patients(F). Relationships for VHL patients are shown by dashed regressionlines, and those for MEN 2 patients are shown by solid regressionlines. Tumor catecholamine content was determined from the productof tumor size and the combined tumor concentration of norepineph-rine and epinephrine. Plasma concentrations of total metanephrinesrepresent the sum of plasma concentrations of normetanephrine andmetanephrine. Plasma concentrations of total catecholamines repre-sent the sum of plasma concentrations of norepinephrine and epi-nephrine. Multiple regression analysis revealed that relative to tumorcatecholamine content, VHL patients had higher (P 5 0.019) plasmaconcentrations of total catecholamines than patients with MEN 2, buthad similar plasma concentrations of total metanephrines.

TABLE 1. Signs and symptoms of pheochromocytoma in VHLdisease and MEN 2

VHL disease MEN 2

Hypertension 6/33 (18) 8/20 (40)Paroxysmal 1/33 (3) 6/20 (30)a

Symptoms 10/33 (30) 12/20 (60)a

Headache 8/33 (24) 7/20 (35)Diaphoresis 6/33 (18) 7/20 (35)Palpitations 6/33 (18) 7/20 (35)Anxiety or panic 5/33 (15) 5/20 (25)

Percentages are given in parentheses.a P , 0.05, a higher incidence in MEN 2 patients than in VHL

patients.

PHEOCHROMOCYTOMA PHENOTYPES 2005

patients differ in their propensity for continuous release ofcatecholamines.

Although numerous studies have reported plasma andurinary catecholamines in VHL patients with pheochromo-cytoma (8, 20, 21), none appears to have noted, as reportedhere, the almost exclusive production of norepinephrine inthis form of hereditary pheochromocytoma. Also, althoughseveral studies have documented that MEN 2 patients withpheochromocytoma often have elevated plasma or urinarylevels of epinephrine (9, 12–14), none has established thatepinephrine production represents a consistent findingamong different kindreds and in large numbers of patientswith MEN 2 and pheochromocytoma.

Previous failure to recognize the clear-cut differences inthe adrenergic and noradrenergic biochemical phenotypes ofpheochromocytomas in MEN 2 and VHL syndrome maystem from two factors. Differences in plasma concentrationsof normetanephrine and metanephrine reflect underlyingdifferences in tumor catecholamine phenotype better than dodifferences in plasma or urinary norepinephrine and epi-nephrine, and advances in molecular genetic diagnosis nowallow unambiguous identification of the underlying muta-tion in the two disorders, whereas previously this was basedlargely on clinical presentation.

The superiority of assays of plasma free metanephrinesover catecholamines for identifying adrenergic or noradren-ergic biochemical phenotypes in pheochromocytoma is il-lustrated by the considerable overlap in plasma concentra-tions of epinephrine and the absence of overlap in plasmaconcentrations of metanephrine among patients with the twohereditary syndromes. Many patients with MEN 2 hadnormal plasma concentrations of epinephrine, and manyVHL patients had normal levels of norepinephrine. In con-trast, all patients with MEN 2 had elevations of plasma meta-nephrine, often accompanied by an increase in normeta-nephrine, and almost all patients with VHL syndrome hadelevations in normetanephrine with little increase in plasmametanephrine. Thus, compared with plasma metanephrines,plasma catecholamines fail to adequately indicate both thepresence and the underlying neurochemical phenotype of apheochromocytoma.

The importance of molecular genetic diagnosis for unam-biguously identifying a mutation of either the RET protoon-cogene or the VHL tumor suppressor gene is illustrated byseveral reports that VHL families with pheochromocytomawere diagnosed erroneously, based on clinical presentation,as having MEN 2 or familial pheochromocytoma (22–25). Asdifferent kindreds can present with different phenotypes, itcan sometimes be difficult to identify the specific hereditarysyndrome, based on clinical presentation alone. In particular,some VHL families have mainly pheochromocytoma, withoccult or delayed manifestations of the syndrome in thecentral nervous system, eye, or other organs (25–27). As MEN2 features a high penetrance of medullary thyroid cancer (28),it is generally less of a problem to identify this syndrome onclinical grounds. Nevertheless, in one kindred of the presentseries, there was no history of either medullary thyroid can-cer or parathyroid disease, and increased plasma concentra-tions of metanephrine provided the only initial clinical ev-idence that the underlying mutation was of the RET generather than of the VHL gene. The above examples illustratehow differences in plasma levels of normetanephrine andmetanephrine can provide a supplementary guide to clinicalpresentation in deciding which gene to test to unambigu-ously identify a particular germline mutation.

The low frequency of hypertension or symptoms foundduring screening of VHL and MEN 2 patients with pheo-chromocytoma agrees with other reports (6–10), reflecting inpart early detection of small tumors that produce insufficientamounts of catecholamines to produce the typical clinicalmanifestations of the tumor. Comparisons with patients withsporadic pheochromocytoma suggest that the small tumorsdetected in MEN 2 families may be more often functionalthan those detected in VHL families (9, 10). This suggestionis supported by the present findings of a higher frequency ofsymptoms and hypertension in MEN 2 than in VHL patientswith pheochromocytoma.

Epinephrine has more potent a- and b2-adrenoceptor-mediated vascular and metabolic effects than norepineph-rine (29, 30). Thus, differences in expression of PNMT andproduction of epinephrine probably contribute to the higherfrequency of signs and symptoms of pheochromocytomas inMEN 2 than in VHL patients. The greater expression of THin pheochromocytomas from MEN 2 than VHL patients is in

FIG. 8. Frequency of hypertension (E and F) or symptoms of pheo-chromocytoma (M and f) as a function of tumor size (A) or plasmaconcentration of total catecholamines (B) in MEN 2 patients (F andf) compared with VHL patients (E and M). The tumor diameterrepresents the average of three dimensions. Plasma concentrations oftotal catecholamines represent the sum of plasma concentrations ofnorepinephrine and epinephrine.

2006 EISENHOFER ET AL. JCE & M • 2001Vol. 86 • No. 5

agreement with previous findings of higher TH activity intumors from MEN 2 patients than in those from other pa-tients (31). The resulting higher rates of catecholamine syn-thesis and releasable tissue stores of catecholamines in pheo-chromocytomas from MEN 2 than from VHL patientsprobably also contribute to the higher frequency of signs andsymptoms of pheochromocytoma in MEN 2 than in VHLpatients.

Despite higher tumor tissue contents of catecholaminesand production of metanephrines, the lower plasma concen-trations of total catecholamines in MEN 2 than in VHL pa-tients indicate that pheochromocytomas in MEN 2 patientsdo not secrete catecholamines as readily or as continuouslyas those in VHL patients. Possibly this difference may reflectthe morphological findings that secretory vesicles in VHLtumor cells are concentrated around cell membranes in ap-parent readiness for release, whereas those in MEN 2 tumorcells are distributed evenly throughout the cytoplasm. Thepossibility that this might result in a more continuous patternof catecholamine release from tumors in VHL patients com-pared with a more episodic pattern in MEN 2 patients isconsistent with previous findings that patients with epineph-rine-secreting tumors present more often with paroxysmalsigns and symptoms than patients with predominantly nor-epinephrine-secreting tumors (2, 5). Greater down-regula-tion of adrenoceptors resulting from continuously higherplasma concentrations of catecholamines in VHL patientscompared with the more intermittently elevated concentra-tions in MEN 2 patients might also contribute to differencesin the clinical presentation of pheochromocytoma among thetwo groups.

Apart from an influence of early tumor detection, lowtissue levels of catecholamines in pheochromocytomas fromVHL patients and low rates of basal catecholamine secretionfrom tumors in MEN 2 patients may be other factors respon-sible for the low frequency of hypertension and symptoms inhereditary pheochromocytoma. Low rates of basal catechol-amine secretion combined with high rates of catecholaminebiosynthesis and consequently metabolism to metanephrinesalso explain the present and previous finding that MEN 2patients with pheochromocytoma have proportionally largerincreases in urinary or plasma metanephrines than in cat-echolamines (12, 32).

The findings of this study that pheochromocytomas inMEN 2 and VHL patients are characterized by distinctlydifferent clinical, biochemical, and morphological featuresraise the question of how RET and VHL gene mutations resultin differences in TH and PNMT expression. As the expres-sions of TH and particularly PNMT are controlled by actionsof steroids on glucocorticoid receptors (33–36), differences inthe expression of these receptors or in local production ofsteroids might account for the differences in catecholamineproduction. Additionally, several other transcription factorsknown to regulate TH and PNMT gene expression (36–40)might be involved in linking RET and VHL gene mutationsto differences in catecholamine synthesis. The above possi-bilities are supported by recent findings that some of thesetranscription factors are differentially expressed in predom-inantly norepinephrine-producing compared with epineph-rine-producing pheochromocytomas (41).

As the adrenal medulla is comprised of subpopulations ofnoradrenergic and adrenergic chromaffin cells (42–44), thedifferent pheochromocytoma tumor cell phenotypes in MEN2 and VHL patients may simply reflect mutation-dependentdevelopment of tumors from different types of chromaffincells. The present and future work aimed at establishing themechanisms that link germline and somatic mutations ofgenes to expression of specific pheochromocytoma tumorcell phenotypes should lead to improved understanding ofthe molecular basis of tumorigenesis, variations in the rate ofdisease progression, tendency to recurrence, metastatic po-tential, and development of novel treatments.

Acknowledgments

The authors are grateful to Ms. Courtney Holmes and Ms. PatriciaSullivan for assistance with catecholamine and metanephrine assays, toDr. Riccardo Gionata Gheri and Ms. Arlene Berman for help with patientstudies, to Ms. Donna Peterson for data management, to Drs. StephenHewitt and Robert Worrell for assistance with procurement of tumortissue, and to the many clinicians and their patients who participated inthe study.

References

1. Kimura N, Miura Y, Nagatsu I, Nagura H. 1992 Catecholamine synthesizingenzymes in 70 cases of functioning and non- functioning phaeochromocytomaand extra-adrenal paraganglioma. Virchows Arch A Pathol Anat Histopathol.421:25–32.

2. Ito Y, Fujimoto Y, Obara T. 1992 The role of epinephrine, norepinephrine, anddopamine in blood pressure disturbances in patients with pheochromocytoma.World J Surg. 16:759–763.

3. Page LB, Raker JW, Berberich FR. 1969 Pheochromocytoma with predominantepinephrine secretion. Am J Med. 47:648–652.

4. Aronoff SL, Passamani E, Borowsky BA, Weiss AN, Roberts R, Cryer PE.1980 Norepinephrine and epinephrine secretion from a clinically epinephrine-secreting pheochromocytoma. Am J Med. 69:321–324.

5. Feldman JM. 1981 Phenylethanolamine-N-methyltransferase activity deter-mines the epinephrine concentration of pheochromocytomas. Res CommunChem Pathol Pharmacol. 34:389–398.

6. Telenius-Berg M, Berg B, Hamberger B, Tibblin S. 1987 Screening for earlyasymptomatic pheochromocytoma in MEN-2. Henry Ford Hosp Med J.35:110–114.

7. Kotzerke J, Stibane C, Dralle H, Wiese H, Burchert W. 1989 Screening forpheochromocytoma in the MEN 2 syndrome. Henry Ford Hosp Med J.37:129–131.

8. Aprill BS, Drake AJ, Lasseter DH, Shakir KM. 1994 Silent adrenal nodulesin von Hippel-Lindau disease suggest pheochromocytoma. Ann Intern Med.120:485–487.

9. Pomares FJ, Canas R, Rodriguez JM, Hernandez AM, Parrilla P, Tebar FJ.1998 Differences between sporadic and multiple endocrine neoplasia type 2Aphaeochromocytoma. Clin Endocrinol (Oxf). 48:195–200.

10. Walther MM, Reiter R, Keiser HR, et al. 1999 Clinical and genetic charac-terization of pheochromocytoma in von Hippel-Lindau families: comparisonwith sporadic pheochromocytoma gives insight into natural history of pheo-chromocytoma. J Urol. 162:659–664.

11. Eisenhofer G, Lenders JW, Linehan WM, Walther MM, Goldstein DS, KeiserHR. 1999 Plasma normetanephrine and metanephrine for detecting pheochro-mocytoma in von Hippel-Lindau disease and multiple endocrine neoplasiatype 2. N Engl J Med. 340:1872–1879.

12. Sato T, Kobayashi K, Miura Y, Sakuma H, Yoshinaga K. 1975 High epi-nephrine content in the adrenal tumors from Sipple’s syndrome. Tohoku J ExpMed. 115:15–19.

13. Hamilton BP, Landsberg L, Levine RJ. 1978 Measurement of urinary epi-nephrine in screening for pheochromocytoma in multiple endocrine neoplasiatype II. Am J Med. 65:1027–1032.

14. Vistelle R, Grulet H, Gibold C, et al. 1991 High permanent plasma adrenalinelevels: a marker of adrenal medullary disease in medullary thyroid carcinoma.Clin Endocrinol (Oxf). 34:133–138.

15. Eisenhofer G, Goldstein DS, Stull R, et al. 1986 Simultaneous liquid-chro-matographic determination of 3,4-dihydroxyphenylglycol, catecholamines,and 3,4-dihydroxyphenylalanine in plasma, and their responses to inhibitionof monoamine oxidase. Clin Chem. 32:2030–2033.

16. Lenders JWM, Eisenhofer G, Armando I, Keiser HR, Goldstein DS, KopinIJ. 1993 Determination of plasma metanephrines by liquid chromatographywith electrochemical detection. Clin Chem. 39:97–103.

PHEOCHROMOCYTOMA PHENOTYPES 2007

17. Hooper D, Kawamura M, Hoffman B, et al. 1997 Tyrosine hydroxylase assayfor detection of low levels of enzyme activity in peripheral tissues. J Chro-matogr B Biomed Appl. 694:317–324.

18. Andrews NC, Faller DV. 1991 A rapid micropreparation technique for ex-traction of DNA-binding proteins from limiting numbers of mammalian cells.Nucleic Acids Res. 19:2499.

19. Heid CA, Stevens J, Livak KJ, Williams PM. 1996 Real time quantitative PCR.Genome Res. 6:986–994.

20. Neumann HP, Berger DP, Sigmund G, et al. 1993 Pheochromocytomas, mul-tiple endocrine neoplasia type 2, and von Hippel-Lindau disease. N EnglJ Med. 329:1531–1538.

21. Atuk NO, Stolle C, Owen Jr JA, Carpenter JT, Vance ML. 1998 Pheochro-mocytoma in von Hippel-Lindau disease: clinical presentation and mutationanalysis in a large, multigenerational kindred. J Clin Endocrinol Metab.83:117–120.

22. Crossey PA, Eng C, Ginalska-Malinowska M, et al. 1995 Molecular geneticdiagnosis of von Hippel-Lindau disease in familial phaeochromocytoma.J Med Genet. 32:885–886.

23. Neumann HP, Eng C, Mulligan LM, et al. 1995 Consequences of direct genetictesting for germline mutations in the clinical management of families withmultiple endocrine neoplasia, type II. JAMA. 274:1149–1151.

24. Gross DJ, Avishai N, Meiner V, Filon D, Zbar B, Abeliovich D. 1996 Familialpheochromocytoma associated with a novel mutation in the von Hippel-Lindau gene. J Clin Endocrinol Metab. 81:147–149.

25. Ritter MM, Frilling A, Crossey PA, et al. 1996 Isolated familial pheochro-mocytoma as a variant of von Hippel-Lindau disease. J Clin Endocrinol Metab.81:1035–1037.

26. Richard S, Beigelman C, Duclos JM, et al. 1994 Pheochromocytoma as the firstmanifestation of von Hippel-Lindau disease. Surgery. 116:1076–1081.

27. Gross DJ, Avishai N, Meiner V, Filon D, Zbar B, Abeliovich D. 1996 Familialpheochromocytoma associated with a novel mutation in the von Hippel-Lindau gene. J Clin Endocrinol Metab. 81:147–149.

28. Howe JR, Norton JA, Wells SA. 1993 Prevalence of pheochromocytoma andhyperparathyroidism in multiple endocrine neoplasia type 2A: results of long-term follow-up. Surgery. 114:1070–1077.

29. Ahlquist RP. 1948 A study of adrenotropic receptors. Am J Physiol.153:586–600.

30. Hjemdahl P, Belfrage E, Daleskog M. 1979 Vascular and metabolic effects ofcirculating epinephrine and norepinephrine. Concentration-effect study indogs. J Clin Invest. 64:1221–1228.

31. Tumer N, Brown JW, Carballeira A, Fishman LM. 1996 Tyrosine hydroxylasegene expression in varying forms of human pheochromocytoma. Life Sci.59:1659–1665.

32. Casanova S, Rosenberg-Bourgin M, Farkas D, et al. 1993 Phaeochromocy-toma in multiple endocrine neoplasia type 2A: survey of 100 cases. ClinEndocrinol (Oxf). 38:531–537.

33. Wurtman RJ, Axelrod J. 1966 Control of enzymatic synthesis of adrenaline inthe adrenal medulla by adrenal cortical steroids. J Biol Chem. 241:2301–2305.

34. Kelner KL, Pollard HB. 1985 Glucocorticoid receptors and regulation of phe-nylethanolamine-N-methyltransferase activity in cultured chromaffin cells.J Neurosci. 5:2161–2168.

35. Fossom LH, Sterling CR, Tank AW. 1992 Regulation of tyrosine hydroxylasegene transcription rate and tyrosine hydroxylase mRNA stability by cyclicAMP and glucocorticoid. Mol Pharmacol. 42:898–908.

36. Wong DL, Siddall BJ, Ebert SN, Bell RA, Her S. 1998 PhenylethanolamineN-methyltransferase gene expression: synergistic activation by Egr-1, AP-2and the glucocorticoid receptor. Brain Res Mol Brain Res. 61:154–161.

37. Ebert SN, Balt SL, Hunter JP, Gashler A, Sukhatme V, Wong DL. 1994 Egr-1activation of rat adrenal phenylethanolamine N-methyltransferase gene. J BiolChem. 269:20885–20898.

38. Her S, Bell RA, Bloom AK, Siddall BJ, Wong DL. 1999 PhenylethanolamineN-methyltransferase gene expression. Sp1 and MAZ potential for tissue-specific expression. J Biol Chem. 274:8698–8707.

39. Ghee M, Baker H, Miller JC, Ziff EB. 1998 AP-1, CREB and CBP transcriptionfactors differentially regulate the tyrosine hydroxylase gene. Brain Res MolBrain Res. 55:101–114.

40. Papanikolaou NA, Sabban EL. 2001 Ability of Egr1 to activate tyrosine hy-droxylase transcription in PC12 cells. Cross-talk with AP-1 factors. J Biol Chem.275:26683–26689.

41. Isobe K, Nakai T, Yashiro T, et al. 2000 Enhanced expression of mRNA codingfor the adrenaline-synthesizing enzyme phenylethanolamine-N-methyl trans-ferase in adrenaline-secreting pheochromocytomas. J Urol. 163:357–362.

42. Michelena P, Moro MA, Castillo CJ, Garcia AG. 1991 Muscarinic receptorsin separate populations of noradrenaline- and adrenaline-containing chro-maffin cells. Biochem Biophys Res Commun. 177:913–919.

43. Vollmer RR, Baruchin A, Kolibal-Pegher SS, Corey SP, Stricker EM, KaplanBB. 1992 Selective activation of norepinephrine- and epinephrine-secretingchromaffin cells in rat adrenal medulla. Am J Physiol. 263:R716–R721.

44. Aunis D, Langley K. 1999 Physiological aspects of exocytosis in chromaffincells of the adrenal medulla. Acta Physiol Scand. 167:89–97.

2008 EISENHOFER ET AL. JCE & M • 2001Vol. 86 • No. 5