PDF hosted at the Radboud Repository of the Radboud University

Nijmegen

The following full text is a publisher's version.

For additional information about this publication click this link.

http://hdl.handle.net/2066/169277

Please be advised that this information was generated on 2017-12-07 and may be subject to

change.

Genotype-specific differences in metabolic profile of pheochromocytoma

and paraganglioma: biomarkers, imaging and metastatic potential

Jyotsna Upendra Rao

Genotype-specific differences in metabolic profile of pheochromocytoma and paraganglioma: biomarkers, imaging and metastatic potential

Colofon

AuthorJyotsna Upendra Rao

Design/lay-out Promotie In Zicht, Arnhem

PrintIpskamp Printing, Enschede

ISBN978-94-9238-042-5

The work presented in this thesis was carried out within the Radboud Institute for Molecular Life Sciences

© 2017 Jyotsna Upendra Rao

All rights reserved. No part of this publication may be reproduced or transmitted in any form by any means, without permission of the author

Proefschrift

ter verkrijging van de graad van doctor aan de Radboud Universiteit Nijmegen

op gezag van de rector magnificus prof. dr. J.H.J.M. van Krieken,volgens besluit van het college van decanen

in het openbaar te verdedigen opdinsdag 23 mei 2017 om 16.30 uur precies

door

Jyotsna Upendra Raogeboren op 1 december 1980

te Udupi, India

Genotype-specific differences in metabolic profile of pheochromocytoma and paraganglioma: biomarkers, imaging

and metastatic potential

PromotorenProf. dr. C.G.J. Sweep Prof. dr. R.A. Wevers

Copromotor Dr. H.J.L.M. Timmers

Manuscriptcommissie Prof. dr. P.M.T. DeenDr. W.P.J. LeendersProf. dr. R.R. de Krijger (Erasmus MC)

Table of Contents

Chapter 1 General Introduction 7

Chapter 2 Genotype-specific abnormalities in mitochondrial function associate with distinct profiles of energy metabolism and catecholamine content in pheochromocytoma and paraganglioma

21

Chapter 3 Genotype–specific differences in the tumour metabolite profile of pheochromocytoma and paraganglioma using untargeted and targeted metabolomics

41

Chapter 4 Overexpression of the natural antisense hypoxia-inducible factor-1α transcript is associated with malignant pheochromocytoma / paraganglioma

65

Chapter 5 Correlation between in vivo 18F-fluorodeoxyglucose PET and immunohistochemical markers of glucose uptake and metabolism in pheochromocytoma and paraganglioma

85

Chapter 6 Discussion 111

Chapter 7 Summary 123

Chapter 8 SamenvattingList of publicationsAcknowledgementsCV

131135139145

General introduction and outline of the thesis

1

9

General Introduction

11.1 Definition, incidence and symptoms

of Pheochromocytoma and Paraganglioma

Pheochromocytoma and paraganglioma (PPGL) are neuroendocrine tumours of neural crest derived cells. As per the recommendation of the World Health Organization, the term pheochromocytoma refers to tumours of sympathetic tissue of adrenal medulla while all other tumours of sympathetic and parasympathetic nervous tissue are called sympathetic and parasympathetic paragangliomas (PGLs), respectively (1). They arise in chromaffin cells of adrenal medulla and extra-adrenal sympathetic and parasympathetic paraganglia. Paraganglia are small organs that mainly consist of neuroendocrine cells derived from embryonic neural crest that have the ability to synthesize and secrete catecholamines (2). In this thesis, pheochromocytoma and sympathetic paraganglioma together, are referred to as PPGLs. PPGLs are further classified by their site viz., pelvis, abdominal and thorax. PPGLs secrete excessive amounts of catecholamines while the para-sympathetic head and neck paragangliomas (HNPGLs) usually do not secrete clinically relevant amounts of catecholamines (1,3). In Western countries, the prevalence of (a history of) PPGL can be estimated at between 1:4500 and 1:1700, with an annual incidence of 3 to 8 cases per 1 million per year in the general population. Undiagnosed lesions account for approximately 5% of adrenal incidentalomas. Clinical symptoms of sympathetic PPGLs are mostly due to excessive catecholamine concentrations in circulation. Symptoms can include paroxysmal episodes of headaches, palpitation and sweating. Additionally pallor, panic/anxiety, fever and nausea might also be seen (1). Most of these PPGLs are benign, however, 10% of pheochromocytomas and 15-35% of paragangliomas show malignancy. Traditionally, malignancy in PPGLs is defined by the presence of metastases of chromaffin tissue in organs like lung, liver, bone and lymph nodes. Prognosis of malignancy in PPGLs is highly variable with a mean 5 year survival rate of 50% (4,5). The outcome worsens with tumour size, organ of metastases and the SDHB genotype (6). Tumours caused by mutations in SDHB and MAX are particularly at high risk for metastases (7). PPGLs are diagnosed by the excessive secretion of catecholamines. Detection of excessive secretion of plasma free metanephrines and urinary excretion of metanephrines are considered the gold standard test. Following biochemical diagnosis, tumour is localized using various imaging methods (8) (see paragraph 1.5).

10

1.2 Genetic background of PPGLs

Nearly 40% of PPGLs are due to hereditary mutations in at least 13 tumour susceptibility genes. These include genes which when mutated result in familial syndromes with autosomal dominant inheritance namely, RET causing Multiple Endocrine Neoplasia type 2A and 2B, VHL causing von Hippel Lindau disease and NF1 causing Neurofibromatosis type 1, all of which mainly present pheochromo-cytomas. Hereditary mutations in fumarate hydratase (FH) gene result in hereditary leiomyomatosis and renal cell cancer and in exceptional cases in paragangliomas (7). Most recently, mutations in the gene for malate dehydrogenase type 2 (MDH2) have been associated with PPGL (9).Genes in which mutations result in familial and non-syndromic presentation include genes encoding succinate dehydrogenase subunits, SDHA, SDHB, SDHC and SDHD and assembly factor SDHAF-2. SDHx-related PPGL are autosomal dominant disorders with usually a high penetrance. Mutations in the tumour suppressor genes TMEM127 and MAX lead to familial pheochromo-cytomas and mutations in the oncogene EPAS1 lead to PPGLs, polycythemia and/or somatostatinoma in patients depending on the degree of mosaicism and the organ affected (7). A summary of the hereditary mutations leading to PPGLs and associated conditions is provided in table 1. Tumours which are not associated with hereditary mutations in any of the known tumour susceptibility genes are known as sporadic tumours. Presence of somatic mutations in RET, NF1, VHL, EPAS1 and HRAS have been reported to cause PPGLs. About 20% of sporadic PPGLs have somatic mutations in NF-1, 14% have mutations in NF1 and VHL and about 5% in HRAS (10-13). Recently, somatic mutations in HRAS, BRAF and TP53 were also identified to occur in PPGLs and it was suggested that HRAS/BRAF mutations could be mutually exclusive (14). It is also important to recognize that even though primary hereditary mutations are present which could lead to cancer, cooperative somatic events are required for initiation and progression of cancer. Studies have also revealed that there is intra- and inter-tumour genetic heterogeneity in tumours from the same patient. Patients with distant metastases showed greater intra-tumour heterogeneity, while local recurrences with metastatic invasion were more closely related to the primary tumour and exhibited a low degree of intra-tumour heterogeneity (12). As PPGLs can arise due to hereditary mutations in a large number of genes, studying these tumours can also have implications on other cancers which share the patho-physiology in terms of mechanisms that lead to tumour formation.

11

General Introduction

1

1.3 Pathophysiology in relation to genetic background of PPGLs

PPGLs can be classified into two clusters based on transcription profiling data. Cluster 1 includes SDHx- and VHL-related tumours. This cluster is characterized by expression of genes involved in hypoxic response. Two recently identified genes associated with PPGLs, FH and EPAS-1, are also classified along with this cluster

Table 1 Hereditary mutations associated with PPGLs and related disorders*

Disease Genes % with PPGL

Main features

Neurofibromatosis type 1

NF1 3 Café-au-lait spots, neurofibromas, axillary and inguinal freckling, Lisch nodules, osseous lesions, optic gliomas, mainly pheochromocytomas

Multiple endocrine neoplasia type 2

RET 6 2A: Medullary thyroid cancer, primary hyperparathyroidism, PPGL2B: Medullary thyroid cancer, PPGL, Marfanoid habitus, mucocutaneous neuromas, gastrointestinal ganglioneuromatosis

von Hippel–Lindau disease

VHL 7 Central nervous system or retinal haemangioblastomas, renal cell carcinoma, PPGL, pancreatic neuroendocrine tumours and cysts, endolymphatic sac tumours, papillary cystadenoma of the epididymis and broad ligament

Hereditary paragangliomas

SDHBSDHDSDHCSDHASDHAF2

1091

<1<0.1

PPGL, rare renal cancers, GISTPPGL, rare renal cancers, GISTPPGL, rare renal cancers, GISTPPGL, GISTHead and neck paraganglioma

Familial pheochromocytomas

TMEM127MAX

11

Mainly pheochromocytomas, rare renal cancersMainly PPGL

Polycythemia paraganglioma syndrome

EPAS1 1 Polycythemia, PPGL, somatostatinoma

Leiomyomatosis and renal cell cancer

FH 1 Cutaneous and uterine leiomyomas, type 2 papillary renal carcinoma, rare PPGL

* adapted from Favier et al., 2015 (7)

12

though originally when the first transcriptomic studies were published they were not known to be associated with the disease. Cluster 2 includes tumours with mutations in RET and NF-1. Most of the sporadic tumours also belong to this category. These tumours show increased activation of the MAPK and PI3K/AKT pathways. TMEM127 and MAX tumours also are identified with cluster 2 (7,15,16). Although, the two clusters seem to activate distinct signalling pathways, they are interconnected in activation of HIF signalling as major driver of tumourigenesis (17). In cluster 1 tumours, HIF-1α and -2α are stabilized either due to accumulation of succinate inhibiting the enzyme prolyl hydroxylase needed for proteolytic degradation of HIF or due to dysfunctional proteosomal HIF degradation complex due to mutation in VHL which is a component of this complex. In both cases, HIF-1α and -2α translocate to the nucleus where, together with the aryl hydrocarbon receptor nuclear translocator, they form an active HIF complex that induces the expression of genes with hypoxia response elements supporting tumour progression. This ‘pseudohypoxic response’ leads to elevated expression of genes involved in glycolysis and angiogenesis. Thus, in cluster 1 tumours, the pseudo -hypoxic drive is hypothesized to mediate an increase in aerobic glycolysis, also known as Warburg effect (18,19). In RET and NF-1 related cluster 2 tumours, HIF activation is driven through activation of Ras/Raf/MAPK pathway. Ras activation results in the activation of downstream PI3K, MAPK and mTORC pathways. Activation of extracellular signal- regulated kinase (ERK), a component of Ras/MAPK pathway leads to phosphorylation of HIF-1α while downstream target of PI3K, nuclear factor kappa B (NF-ϰB) is also known to activate HIF- α pathway. In TMEM127 and MAX related tumours, HIF- α levels are related to mTORC activation (17). A summary of how pathophysiology of cluster 1 and cluster 2 tumours are interrelated through activation of HIF is depicted below in Figure 1.

1.4 Alterations in metabolism in relation to genetic background

Genotype-specific differences have been reported in catecholamine metabolism in PPGLs. Catecholamines are the hormones and neurotransmitters of adreno-medullary, sympathoneuronal and brain catecholaminergic systems. Catecholamines bind to α (α1 and α2) and β (β1, β2 and β3) adrenergic receptors in target cells and eliciting either activation or inhibition of target cells. Physiological effects of catecholamines include increasing the activity of heart muscles, blood vessels, increasing basal metabolic rate, glucose levels and lipolysis.

13

General Introduction

1

The catecholamine biosynthetic pathway is depicted in Figure 1. Catecholamines are synthesized from the precursor amino acid L-tyrosine. Tyrosine is obtained through diet as well as conversion of phenylalanine to tyrosine in liver. Tyrosine is converted to L-dihydroxyphenylalanine (L-DOPA) by the enzyme tyrosine hydroxylase in adrenal chromaffin cells, sympathetic and brain catecholaminergic nerve terminals. L-DOPA is converted to dopamine by a non-specific enzyme L- amino acid decarboxylase (AAD). Dopamine is taken up from the storage vesicles where the enzyme dopamine-β-hydroxylase (DBH) converts dopamine to

Figure 1 Hypoxia inducible factor (HIF) signaling in cluster 1 and cluster 2. Yellow marks highlight the tumour susceptibility genes linked to HIF activation. Adapted from Jochmonova et al., 2013 (17).

14

norepinehrine. Norepinephrine is released into the cytosol by exocytosis where cytoplasmic enzyme phenylethanolamine-N-methyl transferase (PNMT) converts norepinephrine to epinephrine. PNMT is mainly present in adrenal medulla and certain population of chromaffin cells contain PNMT and consequently separate populations contain epinephrine and norepinephrine as their end products (Kvetnansky et al., 2009). Once released, catecholamines are further metabolized or are taken back up into sympathetic nerve endings/chromaffin cells and/or to effector cells. This uptake is mediated by norepinephrine transporter (NET). NET helps in reuptake of norepinephrine and dopamine and has low affinity to epinephrine in neuronal tissue. In extraneuronal tissue, epinephrine is favourably

Figure 2 Pathway for catecholamine biosynthesis and its enzymatic steps. The steps of conversion from L-tyrosine to L-noradrenaline are typical for sympathetic and some brain neurons, and the conversion of L-norepinephrine to L-epinephrine is typical for the adrenal medullary cells and some peripheral and brain neurons. Adapted from Kvetnansky et al., 2009 (20).

15

General Introduction

1taken up over norepinephrine. Reuptake is dependent on Na+/Cl- gradient in neuronal tissue while in extraneuronal chromaffin cells, it is independent of the gradient. It is to be noted that maintenance of Na+/Cl- gradient is an energy dependent process. Vesicular monoamine transporters (VMATs) help in sequestration of catecho-lamines from cytoplasm into chromaffin granules and sympathetic synaptic vesicles. VMAT has two isoforms, VMAT1 and VMAT2. Sympathetic nerves in periphery or brain express VMAT2 while adrenal medullary chromaffin cells express both VMAT1 and 2 (VMAT2 major isoform in humans). The vesicular uptake requires energy which is provided by the proton gradient established by ATP-dependent H+ pump. VMAT is sensitive to the pH gradient and the membrane potential (20). Tumours of different genotypes are characterized by differences in catecholamine biosynthetic and secretory profiles. In contrast to patients with VHL- and SDH-related tumours, PPGLs with mutations in RET and NF1 present with tumours characterized by increased plasma metanephrine concentrations indicative of increased epinephrine production. On the other hand, PPGLs with mutations in VHL showed increased production of norepinephrine as evidenced by increased plasma normetanephrine levels (21). Patients with SDHB- and SDHD-related tumours showed increased plasma levels of methoxytyramine indicative of increased dopamine production. However, overall tumour catecholamine content is lower in cluster 1 tumours, compared to cluster 2 tumours. In addition, RET- and NF1-related PPGLs and sporadic tumours with similar secretory profiles showed a lower rate of catecholamine secretion than VHL-, SDH-related and norepinephrine secreting sporadic tumours which did not have high epinephrine content (22). Other aspects of metabolism altered in a genotype-specific way include oxidative phosphorylation and glycolysis. Mitochondrial oxidative phosphorylation plays a pivotal role in cellular energy production through activity of respiratory chain enzyme complexes. Of the four respiratory chain enzyme complexes (complex I-IV), complex II or succinate dehydrogenase has a unique role as it is an important component of the respiratory chain as well as the Krebs cycle. The mitochondrial succinate dehydrogenase (SDH) complex catalyzes the oxidation of succinate to fumarate in the Krebs cycle, and feeds electrons to the respiratory chain ubiquinone pool. It consists of four nuclear encoded subunits A, B, C and D. The SDHA subunit contains the covalently attached flavin adenine dinucleotide (FAD) cofactor and the succinate binding site. SDHB subunit contains the iron-sulphur clusters and SDHC and D subunits are involved in membrane anchoring and electron transfer to ubiquinone. Thus, mutation in any of the SDH subunits is expected to impair the enzyme activity leading to succinate accumulation as well as electron transfer function. As explained in section 1.3,

16

consistent with the hypothesis of activation of pseudohypoxic response, increased levels of HIF-2α are observed in SDH- and VHL-related tumours when compared to cluster 2 tumours (16,23). Increased HIFα levels and nuclear translocation leads to increased expression of a wide variety of genes involved in angiogenesis, glycolysis, invasion and metastasis. Increased vascular density as evidenced by increased CD34 staining is observed in SDH- and VHL- related tumours. VHL-related tumours also exhibit increased lactate dehydrogenase activity (16). However, the impact of genotype on tumour metabolism in PPGLs had not been systematically studied before the studies that comprise this thesis were initiated.

1.5 Role of metabolism in imaging of PPGLs

Various functional imaging modalities are used in localization of PPGLs. These functional imaging contrast agents target catecholamine metabolism (synthesis, storage and secretion) or energy metabolism (glycolysis). 123/131I-metaiodobenzyl-guanidine (123/131I-MIBG) scintigraphy and 18F- fluorodopamine positron emission tomography (18F-FDA PET) target the norepinephrine transporter of the PPGL cell membrane and the vesicular monoamine transporters in the membrane of intra cellular vesicles. These transporters facilitate the re-uptake and storage of catecholamines, respectively. Thus, they are effective in detecting RET- and NF-1-related tumours (24). On the other hand, 18F- fluorodeoxy glucose (18F-FDG)-PET has high sensitivity and specificity in detecting VHL- and SDH-related tumours including metastatic lesions (24,25). The avidity of cluster 1 tumours in uptake of 18F-FDG is hypothesized to be reflective of increased glycolysis which is related to the pseudohypoxic drive in these tumours. The relationship between tumour energy metabolism and functional imaging characteristics is the topic of investigation in this thesis. The findings hold promise to improve the imaging tools used for tumour localization.

1.6 Aims and scope of the thesis

The aim of this thesis is to study the impact of genotype-specific alterations in tumour energy metabolism in PPGLs on diagnostic characteristics and metastatic potential. For this purpose, targeted and untargeted ex-vivo metabolomic investigations of PPGL tumour tissues are compared with clinical characteristics and in vivo functional imaging of patients. In chapter 1, we describe the impact of genotype specific alterations in respiratory chain enzyme activities on energy and catecholamine metabolism in PPGLs. For this purpose, we use respiratory

17

General Introduction

1chain enzyme assays to assess the respiratory chain activity and proton nuclear magnetic resonance (1H NMR) spectroscopy to analyze metabolite levels. In chapter 2, we further extend the study described in chapter 1 to understand the genotype-specific differences in the PPGL metabolic profile. Towards this end, untargeted and targeted metabolic profiling is carried out using 1H NMR spectroscopy and HPLC tandem mass spectrometry. We investigate the link between altered energy metabolism and purine and amino acid metabolism.In chapter 3, we investigate gene expression changes that could help PPGLs to cope with prolonged pseudohypoxia. Early hypoxic response leads to activation of p53 through HIF-1α mediated stabilization. However, in PPGLs, loss of p53 is associated with sustained pseudohypoxia (16). Here we investigate the role of antisense HIF, which encodes the antisense template of the 3’-untranslated region of HIF-1α mRNA and inhibits the translation of HIF-1α mRNA after chronic hypoxia. In chapter 4, we investigate how the genotype-specific differences in PPGL energy metabolism are related to increased 18F-FDG uptake observed in-vivo in PPGLs of various genotypes. In chapter 5, the findings presented in this thesis, future perspectives for application of findings and further research questions that could be addressed are discussed. Chapter 6 presents a summary of the findings of the research work of this thesis.

18

1.7 References1. Lenders JW, Eisenhofer G, Mannelli M, Pacak K. Phaeochromocytoma. Lancet. 2005;366(9486):665-675.2. Welander J, Soderkvist P, Gimm O. Genetics and clinical characteristics of hereditary pheochromocytomas

and paragangliomas. Endocrine-related cancer. 2011;18(6):R253-276.3. van Duinen N, Corssmit EP, de Jong WH, Brookman D, Kema IP, Romijn JA. Plasma levels of free

metanephrines and 3-methoxytyramine indicate a higher number of biochemically active HNPGL than 24-h urinary excretion rates of catecholamines and metabolites. Eur J Endocrinol. 2013;169(3):377-382.

4. Harari A, Inabnet WB, 3rd. Malignant pheochromocytoma: a review. Am J Surg. 2011;201(5):700-708.5. Parenti G, Zampetti B, Rapizzi E, Ercolino T, Giache V, Mannelli M. Updated and new perspectives on

diagnosis, prognosis, and therapy of malignant pheochromocytoma/paraganglioma. J Oncol. 2012;2012:872713.

6. Amar L, Baudin E, Burnichon N, Peyrard S, Silvera S, Bertherat J, Bertagna X, Schlumberger M, Jeunemaitre X, Gimenez-Roqueplo AP, Plouin PF. Succinate dehydrogenase B gene mutations predict survival in patients with malignant pheochromocytomas or paragangliomas. J Clin Endocrinol Metab. 2007;92(10):3822-3828.

7. Favier J, Amar L, Gimenez-Roqueplo AP. Paraganglioma and phaeochromocytoma: from genetics to personalized medicine. Nat Rev Endocrinol. 2015;11(2):101-111.

8. van Berkel A, Lenders JW, Timmers HJ. Diagnosis of endocrine disease: Biochemical diagnosis of phaeochromocytoma and paraganglioma. Eur J Endocrinol. 2014;170(3):R109-119.

9. Cascon A, Comino-Mendez I, Curras-Freixes M, de Cubas AA, Contreras L, Richter S, Peitzsch M, Mancikova V, Inglada-Perez L, Perez-Barrios A, Calatayud M, Azriel S, Villar-Vicente R, Aller J, Setien F, Moran S, Garcia JF, Rio-Machin A, Leton R, Gomez-Grana A, Apellaniz-Ruiz M, Roncador G, Esteller M, Rodriguez-Antona C, Satrustegui J, Eisenhofer G, Urioste M, Robledo M. Whole-exome sequencing identifies MDH2 as a new familial paraganglioma gene. Journal of the National Cancer Institute. 2015;107(5).

10. Burnichon N, Buffet A, Parfait B, Letouze E, Laurendeau I, Loriot C, Pasmant E, Abermil N, Valeyrie-Al-lanore L, Bertherat J, Amar L, Vidaud D, Favier J, Gimenez-Roqueplo AP. Somatic NF1 inactivation is a frequent event in sporadic pheochromocytoma. Hum Mol Genet. 2012;21(26):5397-5405.

11. Burnichon N, Vescovo L, Amar L, Libe R, de Reynies A, Venisse A, Jouanno E, Laurendeau I, Parfait B, Bertherat J, Plouin PF, Jeunemaitre X, Favier J, Gimenez-Roqueplo AP. Integrative genomic analysis reveals somatic mutations in pheochromocytoma and paraganglioma. Hum Mol Genet. 2011;20(20):3974-3985.

12. Crona J, Backman S, Maharjan R, Mayrhofer M, Stalberg P, Isaksson A, Hellman P, Bjorklund P. Spatiotemporal Heterogeneity Characterizes the Genetic Landscape of Pheochromocytoma and Defines Early Events in Tumourigenesis. Clin Cancer Res. 2015;21(19):4451-4460.

13. Zhuang Z, Yang C, Lorenzo F, Merino M, Fojo T, Kebebew E, Popovic V, Stratakis CA, Prchal JT, Pacak K. Somatic HIF2A gain-of-function mutations in paraganglioma with polycythemia. N Engl J Med. 2012;367(10):922-930.

14. Luchetti A, Walsh D, Rodger F, Clark G, Martin T, Irving R, Sanna M, Yao M, Robledo M, Neumann HP, Woodward ER, Latif F, Abbs S, Martin H, Maher ER. Profiling of somatic mutations in phaeochromo-cytoma and paraganglioma by targeted next generation sequencing analysis. Int J Endocrinol. 2015;2015:138573.

15. Dahia PL. Transcription association of VHL and SDH mutations link hypoxia and oxidoreductase signals in pheochromocytomas. Annals of the New York Academy of Sciences. 2006;1073:208-220.

16. Favier J, Briere JJ, Burnichon N, Riviere J, Vescovo L, Benit P, Giscos-Douriez I, De Reynies A, Bertherat J, Badoual C, Tissier F, Amar L, Libe R, Plouin PF, Jeunemaitre X, Rustin P, Gimenez-Roqueplo AP. The Warburg effect is genetically determined in inherited pheochromocytomas. PloS one. 2009;4(9):e7094.

17. Jochmanova I, Yang C, Zhuang Z, Pacak K. Hypoxia-inducible factor signaling in pheochromocytoma: turning the rudder in the right direction. Journal of the National Cancer Institute. 2013;105(17):1270-1283.

19

General Introduction

118. Lee S, Nakamura E, Yang H, Wei W, Linggi MS, Sajan MP, Farese RV, Freeman RS, Carter BD, Kaelin WG,

Jr., Schlisio S. Neuronal apoptosis linked to EglN3 prolyl hydroxylase and familial pheochromocytoma genes: developmental culling and cancer. Cancer cell. 2005;8(2):155-167.

19. Selak MA, Armour SM, MacKenzie ED, Boulahbel H, Watson DG, Mansfield KD, Pan Y, Simon MC, Thompson CB, Gottlieb E. Succinate links TCA cycle dysfunction to oncogenesis by inhibiting HIF-alpha prolyl hydroxylase. Cancer cell. 2005;7(1):77-85.

20. Kvetnansky R, Sabban EL, Palkovits M. Catecholaminergic systems in stress: structural and molecular genetic approaches. Physiological reviews. 2009;89(2):535-606.

21. Eisenhofer G, Lenders JW, Timmers H, Mannelli M, Grebe SK, Hofbauer LC, Bornstein SR, Tiebel O, Adams K, Bratslavsky G, Linehan WM, Pacak K. Measurements of plasma methoxytyramine, norme-tanephrine, and metanephrine as discriminators of different hereditary forms of pheochromocytoma. Clin Chem. 2011;57(3):411-420.

22. Eisenhofer G, Pacak K, Huynh TT, Qin N, Bratslavsky G, Linehan WM, Mannelli M, Friberg P, Grebe SK, Timmers HJ, Bornstein SR, Lenders JW. Catecholamine metabolomic and secretory phenotypes in phaeochromocytoma. Endocr Relat Cancer. 2011;18(1):97-111.

23. Pollard PJ, Briere JJ, Alam NA, Barwell J, Barclay E, Wortham NC, Hunt T, Mitchell M, Olpin S, Moat SJ, Hargreaves IP, Heales SJ, Chung YL, Griffiths JR, Dalgleish A, McGrath JA, Gleeson MJ, Hodgson SV, Poulsom R, Rustin P, Tomlinson IP. Accumulation of Krebs cycle intermediates and over-expression of HIF1alpha in tumours which result from germline FH and SDH mutations. Human molecular genetics. 2005;14(15):2231-2239.

24. Timmers HJ, Taieb D, Pacak K. Current and future anatomical and functional imaging approaches to pheochromocytoma and paraganglioma. Horm Metab Res. 2012;44(5):367-372.

25. Timmers HJ, Chen CC, Carrasquillo JA, Whatley M, Ling A, Havekes B, Eisenhofer G, Martiniova L, Adams KT, Pacak K. Comparison of 18F-fluoro-L-DOPA, 18F-fluoro-deoxyglucose, and 18F-fluorodopamine PET and 123I-MIBG scintigraphy in the localization of pheochromocytoma and paraganglioma. The Journal of clinical endocrinology and metabolism. 2009;94(12):4757-4767.

Genotype-specific abnormalities in mitochondrial function associate with distinct profiles of energy metabolism and catecholamine content in pheochromocytoma and paraganglioma

Rao JU, Engelke UFH, Rodenburg RJT, Wevers RA, Pacak K, Eisenhofer G, Qin N, Kusters B, Goudswaard AG, Lenders JWM, Hermus ARMM, Mensenkamp AR, Kunst HPM, Sweep FCGJ, Timmers HJLM.

Clin Cancer Res. 2013 19(14):3787-95.

2

22

Abstract

Purpose: Pheochromocytomas and paragangliomas (PGLs) are neuroendocrine tumours of sympathetic and parasympathetic paraganglia. The present study investigated the relationships between genotype-specific differences in mitochondrial function and catecholamine content in PGL tumours. Experimental Design: Respiratory chain enzyme assays and 1H-NMR spectroscopy at 500 MHz, were performed on homogenates of 35 sporadic PGLs and 59 PGLs from patients with hereditary mutations in SDHB, SDHD, SDHAF-2, VHL, RET, NF1 and MAX. Results: In SDHx related PGLs, a significant decrease in complex II activity (p<0.0001) and a significant increase in complex I, III and IV enzyme activities were observed when compared to sporadic, RET and NF1 tumours. Also, a significant increase in citrate synthase (p<0.0001) enzyme activity was observed in SDHx related PGLs when compared to sporadic, VHL, RET and NF1 related ones. An increase in succinate accumulation (p<0.001) and decrease in ATP/ADP/AMP accumulation (p<0.001) was observed when compared to sporadic PGLs and PGLs of other genotypes. Positive correlations (p<0.01) were observed between respiratory chain complex II activity and total catecholamine content and ATP/ADP/AMP and total catecholamine contents in tumour tissues. Conclusions: The present study for the first time establishes relationship between determinants of energy metabolism like activity of respiratory chain enzyme complex II, ATP/ADP/AMP content and catecholamine content in PGL tumours. Also, the present study for the first time successfully uses NMR spectroscopy to detect catecholamines in PGL tumours and provides ex vivo evidence for the accumulation of succinate in PGL tumours with a SDHx mutation.

23

Energy Metabolism and Catecholamine Content in PGL

2

2.1 Introduction

Pheochromocytomas and paragangliomas are neuroendocrine tumours of sympathetic (PGLs) and parasympathetic paraganglia. PGLs of sympathetic origin (adrenal medulla and extra adrenal sympathetic tissue of abdomen, pelvis and chest) usually produce catecholamines whereas the tumours of parasym pathetic origin (head and neck PGLs (HNPGLs)) usually do not produce significant amounts of catecholamines (1). At least 30%-35% of the PGLs are caused by germline mutations of ten identified tumour susceptibility genes (2). These include VHL (von Hippel-Lindau), RET (rearranged during transfection), NF1 (neuro fibromatosis type 1), SDHA/B/C/D (succinate dehydrogenase subunits A, B, C and D), SDHAF2 (succinate dehydrogenase assembly factor 2) and the more recently reported TMEM127 (transmembrane protein 127) and MAX (myc-associated factor X) (2). In 17% of sporadic tumours, somatic mutations in RET, VHL, MAX and more recently, HIF-2α and NF1 have been reported (3-5). Based on transcriptional profiling studies, PGLs can be classified into two clusters: cluster 1 and cluster 2 (6, 7). Cluster 1 tumours (VHL, SDHA/B/C/D/AF2) are characterized by increased expression of genes involved in (pseudo)hypoxia, cell proliferation, angiogenesis, electron transport chain and the Krebs cycle and abnormal function of oxidoreductases. Cluster 2 tumours (RET, NF1) show an increased expression of genes involved in protein synthesis, kinase signaling, endocytosis and maintenance of a differentiated chomaffin cell catecholamine biosynthetic and secretory phenotype. Sporadic PGLs are distributed between the two major clusters based on their gene expression pattern and catecholamine phenotype (6). SDH is an important component of the mitochondrial electron transport chain. In tumours with SDHx mutations, the ability of cells for oxidative phos-phorylation is compromised (7-10). Also, it has been demonstrated in vitro that accumulation of succinate in cells silenced for SDH causes inhibition of prolyl hydroxylase activity resulting in stabilization of hypoxia-inducible factors (HIF) -1α and -2α (11, 12). HIF-1α and -2α then translocate to the nucleus where, together with aryl hydrocarbon receptor nuclear translocator (ARNT), they form an active HIF complex that induces the expression of genes with hypoxia response elements that support tumour progression via different signaling pathways. Thus, in cluster 1 tumours, the pseudo-hypoxic drive is hypothesized to mediate an increase in aerobic glycolysis, also known as Warburg effect. This is supported by increased HIF-α protein level combined with lower SDH activity and increased glycolysis as indicated by lactate dehydrogenase activity (7). The differences between cluster 1 and 2 tumours are also characterized by differences in catecholamine biosynthetic and secretory profiles (13). PGLs with

24

mutations in RET and NF1 produce both epinephrine and norepinephrine and have low rate constants for catecholamine secretion, whereas SDHx and VHL–related tumours mainly produce norepinephrine and have high rate constants for catecholamine secretion. Tumour catecholamine content is lower in cluster 1 tumours, compared to cluster 2 tumours. Also, it is well known that sequestration of catecholamines into chromaffin granules through vesicular monoamine transporters (VMAT) and re-uptake of catecholamines via norepinephrine transporter (NET) are active energy dependent processes. Differences in catecholamine phenotypes may thus in part be explained by mutation-dependent changes in energy metabolism. In the present study, we therefore investigated relationships between geno-type-specific differences in mitochondrial function and catecholamine content in PGL tumours. Ninety PGL tissues of various genotypes were included. Besides functional assays for respiratory complex I-IV and citrate synthase (CS), 1H nuclear magnetic resonance (NMR) spectroscopy was performed to provide an overview of the intracellular metabolome with specific focus on catecholamines, AMP/ADP/ATP and intermediates of glycolysis and Krebs cycle.

2.2 Materials and methods

Patients Patients with histologically proven PGLs evaluated at the department of General Internal Medicine, division of Endocrinology of the Radboud University Nijmegen Medical Centre (RUNMC), Nijmegen, the Netherlands and at Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD), National Institute of Health (NIH), Bethesda, MD, USA were considered for the study. Tumour tissue samples of consecutive patients from RUNMC who underwent surgical resection between 1988 and 2012 were included in the study. NIH patients underwent surgery between 2003 and 2010. Frozen primary tumour tissues from 64 patients at RUNMC and 26 patients at NICHD were included. The presence of germline mutations and large deletions in SDHB/C/D, RET, VHL and -since 2011- in SDHA, SDHAF2, TMEM127 and MAX, was investigated using standard procedures. Data were collected under conditions of regular clinical care, with ethical committee approval obtained for the use of those data for scientific purposes at RUNMC. The study was approved by the Institutional Review Board of NICHD, and all patients gave written informed consent before testing. The details of the patients’ clinical characteristics and genotype are listed in Table 1.

25

Energy Metabolism and Catecholamine Content in PGL

2

Tumour tissue processing Tumour tissues resected from the patients described above were procured as early as possible, the dimensions of the tumour were recorded by the pathologist and a small piece of the tumour tissue was weighed and snap frozen and stored in liquid Nitrogen and later used for experimental purposes. For histological confirmation, additional slices were stained with hematoxylin and eosin and re-evaluated by an independent pathologist (BK).

Respiratory chain enzyme assays Frozen tumour specimens (~40 mg) were homogenized on melting ice in sucrose- EDTA-phosphate buffer (0.25 M sucrose, 2 mM EDTA, 10 mM K2PO4, pH 7.4, 8% w/v) using a hand held glass/glass homogenizer. Homogenates were subsequently centrifuged at 600xg at 4oC for 10 min and supernatants used for the determination of the activities of respiratory chain enzyme complexes I, II, III and IV and the mitochondrial matrix enzyme, CS. These assays measured the formation of a spectrophotometrically detectable end-product at regular time intervals and were performed on Konelab 20XT clinical chemistry analyzer (Thermo Scientific, Finland) as described elsewhere (14). The protein concentrations in the super- natants were also measured in parallel using pyrogallol red-molybdate complex method as described earlier (15). The enzyme activities were normalized to mg protein. Five samples were excluded from analysis (Sporadic- 2, SDHB, VHL and NF1- 1 each) as the tumour tissue homogenate contained blood which could affect determinations and interpretations of respiratory chain enzyme activities and protein concentrations.

Table 1 Patient Characteristics

Genotype N (patients) Agey (mean±SD)

Gender (M/F)

N (tumours)

Tumour Location (A/E/HN)

Tumour volume (cm3)

Sporadic 35 47.5 ± 14.8 18/17 35 32/3/0 218.6 ± 452.5SDHB 13 32 ± 11.9 10/3 15 1/14/0 273.1 ± 316.2SDHD 8 44 ± 10.4 6/2 8 1/1/6 27.9 ± 47SDHAF2 1 24 1/0 1 0/0/1 18.9VHL 9 30.7 ± 13 6/3 9 7/2/0 61.5 ± 84.9MEN-2 13 37.9 ± 12.8 6/7 15 15/0/0 93.4 ± 171.6NF1 8 43.2 ± 17.4 5/3 8 8/0/0 107.6 ± 80.6MAX 3 48 ± 14 2/1 3 3/0/0 63.2 ± 71.9

N: No. of patients, y: years, M: Male, F: Female, A: Adrenal, E: Extraadrenal, HN: head and neck

26

1H NMR Spectroscopy 1H NMR spectroscopy was performed in frozen tumour tissues to determine concentrations of intermediates of energy metabolism (Krebs cycle and glycolysis), catecholamines and their metabolites. One sporadic, 2 SDHB, 1 VHL and 2 MAX tumours were excluded from the experiment as the amount of starting material was low. The tumour tissues were homogenized on ice in 10% w/v of distilled water using a hand held Teflon/glass homogenizer. The samples were then centrifuged at 16,000xg for 10 min at 4oC and the supernatants were subjected to ultrafiltration with Vivaspin Turbo 15, 10 kDa filters (Sartorius, Germany). The ultrafiltrates were diluted with water to 700 µl, pH was adjusted to 2.5 and 20 µl of 20.2 mM sodium 3-trimethylsilyl-2,2,3,3-tetradeuteropropionate (TSP) in D2O was added to the samples. The samples were then placed in 5 mm NMR tubes and 1H NMR spectra were obtained using a Bruker 500 MHz spectrometer (pulse angle 90o, 7 µs pulses with a delay time of 4 s, number of scans 256). The water resonance was suppressed by gated irradiation centered on water frequency (16, 17). Concentrations of succinate in the tumour tissues were estimated by integrating the area under the peak at 2.66 ppm. Differences in peak heights were clearly observed for tumours with high, absent and low levels of succinate (Supplementary Figure 1). Tumour tissue ATP/ADP/AMP content were estimated by integrating the area under peaks in the region 6.18-6.21 ppm. Since each of the three compounds differ by the presence of one phosphate group, spectral peaks for these metabolites cannot be distinguished using 1H NMR spectroscopy. Epinephrine content was estimated by integrating area under the peaks for the triplet at 2.73 ppm and the total catecholamine content was estimated by integrating the area under the peaks for multiplets in the region 6.86-6.98 ppm (Supplementary Figure 1). Further, norepinephrine content was estimated by calculating the difference between total catecholamine and epinephrine contents. Other amines, which could contribute to peaks in the region 6.86-6.98 ppm are dopamine and 3-methoxytyramine. However, the concentration of these compounds as measured by HPLC is in the nanomolar range (18), much lower than the detection limit of 1H NMR spectroscopy, which is in micromolar range. The catecholamine content of the tumour tissue as estimated by 1H NMR spectroscopy was validated in a small subset of 22 samples using HPLC. For this purpose, the frozen tumour tissues (~5 mg) were weighed accurately, transferred to a processing tube and 5 volumes of 0.4 M perchloric acid containing 0.5 mM EDTA is added and homogenized on ice using a homogenizer (Polytron). The tubes are then spun for 15 minutes at 3000 rpm in a refrigerated centrifuge to separate the precipitated proteins and cell debris. The perchloric acid extract is separated from the pellet, frozen on dry ice and stored at -800C until assayed for catecholamines. Further, the perchloric acid extract is prepared using alumina extraction method as

27

Energy Metabolism and Catecholamine Content in PGL

2

described previously (19). A significant correlation and linear relationship was observed between the two methods (Supplementary Figure 2).

Statistical Methods Statistical analyses were performed using SPSS (SPSS Inc., v.18) and GraphPad Prism 6 software (GraphPad, La Jolla, CA). The data was analyzed using independent samples Kruskal-Wallis test and Dunn’s post-test was used to compare the different genotypes and adjusted P values were reported. Correlation between respiratory chain enzyme activities and tumour tissue catecholamine content was examined using Spearman’s correlation test. Statistical significance was accepted at P value <0.05. Further, estimation of catecholamine content by 1H NMR spectroscopy and HPLC and comparison of respiratory chain enzyme activities with total catecholamine content was carried out using Passing-Bablok regression analysis (20).

2.3 Results

Respiratory chain enzyme activities The activity of the respiratory chain complex II was deficient in all SDHx tumours as expected (Fig 1B; p<0.0001). The activity of the complexes I, III and IV was significantly higher in SDHx tumours than in VHL, RET, NF1 and sporadic tumours (Fig 1A, C, D, p<0.05). This was even more clearly so for citrate synthase activity (Fig 1E; p<0.0001). The VHL tumours showed a lower activity for Complex I when compared to sporadic tumours (Fig 1A, Table 2, p<0.05) and complex III when compared to sporadic (Fig 1C, Table 2, p<0.01) and cluster II tumours (Fig 1C, Table 2, p<0.05). In contrast, the MAX tumour group had a tendency of higher activity for the complexes II, III and IV when compared to the other genotypes (Fig 1B-D). Further, for the various SDH mutations no mutation specific differences could be observed in the activities of respiratory chain enzyme complexes I-IV. Although a low complex II activity was observed in SDHB related tumours when compared to SDHD related ones, the sensitivity of the assay at such low enzyme activity levels precludes such an analysis and interpretation of the results (supplementary figure 3).

Detection of energy metabolism intermediates and tumour tissue catecholamines using 1H-NMR spectroscopy The NMR spectra of the tumour homogenates showed very high succinate (p<0.001) levels in all SDHx cases as expected (Fig 2A). Mutation specific differences in the levels of succinate could not be observed for the tumours with various SDH

28

mutations (supplementary figure 3). Succinate was not NMR detectable or very low in all other tumour samples except for 1 tumour in the sporadic group. Citrate was present in high concentration in four sporadic tumours. Low concentrations of pyruvate, without genotype specific differences, were observed in all PGL tissues (data not shown).

Figure 1 Respiratory chain enzyme activity in PGL tumour tissues of different genotypes.A-E. Dot plots depicting the respiratory chain enzyme activities of respiratory chain enzyme complexes I to IV and Citrate Synthase (CS) across different genotypes in mU normalized to mg protein concentrations. Horizontal line represents the mean. Data sets having different alphabets above them are significantly different (P<0.05).

A

C

E

B

D

29

Energy Metabolism and Catecholamine Content in PGL

2

The differences in high energy phosphate content between the tumours were striking. Proton NMR spectroscopy cannot discriminate between ATP, ADP and AMP and therefore figure 2B shows the sum of the three high energy phosphates. The concentration of ATP/ADP/AMP was consistently very low (p<0.0001) in all SDHx tumours. A very low content also occurred in the other tumour groups but the ATP/ADP/AMP concentration was rather variable in these groups. RET tumours had high ATP/ADP/AMP when compared to sporadic (p<0.01) and SDHx tumours (p<0.0001).

Epinephrine was proton NMR-undetectable in all SDHx and VHL tumours. The RET tumours showed high epinephrine concentrations when compared to sporadic (p<0.05), SDHx (p<0.0001) and VHL (p<0.001) tumours (Fig 2C). In SDHx tumours, norepinephrine was also very low when compared to sporadic (p<0.01), RET (p<0.0001) and NF1 tumours (p< 0.05) while 50% of the VHL tumours produced significant amounts of norepinephrine (Fig 2D). Total catecholamine peaks were below detection limit in 33 tumour samples. Seven of these samples were subjected to HPLC which detected the presence of catecholamines in these tissues.

Correlation of energy metabolism with tumour tissue catecholamine contentPositive correlations (p<0.001) were observed between activities of respiratory chain complex II and concentrations of epinephrine, norepinephrine and total catecholamine content in tumour tissues. ATP/ADP/AMP content of PGL tumours also showed a positive correlation with tumour tissue epinephrine, norepinephrine

Table 2 Comparison of respiratory chain complex activities between different genotypes

Genotype SDHx vs VHL vs

Sporadic MEN-2 and NF1 Sporadic MEN-2 and NF1Complex I 0.3164 0.0278 0.0478 0.4221

Complex II < 0.0001 < 0.0001 0.1150 0.1703

Complex III 0.0050 0.0165 0.0082 0.0137

Complex IV 0.0397 0.0439 0.0600 0.0733

Note: Listed in the table are P values (adjusted) for the various comparisons. Highlighted in bold letters are comparisons which attained statistical significance.

30

and total catecholamine levels (Table 3). Parasympathetic PGLs were excluded from this analysis as they do not produce catecholamines. Further, Passing-Bablok regression statistics for comparison between complex II activity, tumour ATP/ADP/AMP content and total catecholamine content demonstrated a linear relationship (Figure 3).

Figure 2 Accumulation of intermediates of energy metabolism and catecholamine metabolism in PGL tumour tissues of different genotypes as determined by 1H-NMR spectroscopy. A. Dot plot depicting the tumour tissue succinate concentrations expressed as nmol per mg tumour tissue across different genotypes. B. Dot plot depicting the tumour tissue ATP/ADP/AMP concentrations expressed as mmol per mg tumour tissue across different genotypes. C. Dot plot representing the tumour tissue epinephrine concentrations expressed as µmol per mg tumour tissue across different genotypes. D. Dot plot representing the tumour tissue norepinephrine concentrations expressed as µmol per mg tumour tissue across different genotypes. Horizontal line represents the mean. Data sets having different alphabets above them are significantly different (P<0.05). In 13 tumour samples, succinate and ATP/ADP/AMP and in 33 tumour samples total catecholamine peaks were below detection limit. They have been represented as half the lowest detectable value.

A

B

C

D

31

Energy Metabolism and Catecholamine Content in PGL

2

Tabl

e 3

Cor

rela

tion

of r

espi

rato

ry c

hain

enz

yme

acti

vity

wit

h tu

mou

r tis

sue

cate

chol

amin

e co

nten

t

Cate

chol

amin

esCI

(m

U/m

g pr

otei

n)

CII

(mU

/mg

prot

ein)

CIII

(m

U/m

g pr

otei

n)

CIV

(m

U/m

g pr

otei

n)

ATP/

AD

P/A

MP

(mm

ol/m

g ti

ssue

)

Spea

rman

’s rh

oE

(nm

ol/m

g tis

sue)

Corr

elat

ion

Coef

ficie

nt0.

226

0.42

30.

182

0.00

60.

366

Sig.

(2-t

aile

d)0.

046

0.00

010.

097

0.95

50.

0001

N78

7878

7882

NE

(nm

ol/m

g tis

sue)

Corr

elat

ion

Coef

ficie

nt0.

160

0.47

90.

078

0.06

60.

390

Sig.

(2-t

aile

d)0.

163

0.00

010.

495

0.56

90.

0001

N78

7878

7882

Tota

l cat

echo

lam

ines

(n

mol

/mg

tissu

e)Co

rrel

atio

n Co

effic

ient

0.20

80.

517

0.14

80.

033

0.45

8Si

g. (2

-tai

led)

0.06

70.

0001

0.19

60.

774

0.00

01N

7878

7878

82

E: e

pine

phri

ne, C

I, II,

III a

nd IV

: res

pira

tory

cha

in e

nzym

e co

mpl

exes

I, II

, III

and

IV. H

ighl

ight

ed in

bol

d le

tter

s ar

e co

mpa

riso

ns w

hich

att

aine

d st

atis

tica

l si

gnifi

canc

e.

32

2.4 Discussion

The present study establishes differences in mitochondrial energy pathways in metabolic processes in PGLs and provides novel insight into how deregulation of energy metabolism might impact catecholamine phenotypic features of cluster 1 and cluster 2 tumours. Further, the study also provides ex vivo evidence for the accumulation of succinate in SDH-related tumours and successfully uses 1H-NMR spectroscopy for detection of catecholamines in PGL tumour tissues. In accordance with previous reports (7-10), we observed that the activity of SDH or respiratory chain enzyme complex II is low in SDH-related tumours. Interestingly, reduction of complex II activity in SDH-related tumours was associated with increased activities of other respiratory chain complexes I, III and IV and CS. The increased CS activity also indicates an increase in the mitochondrial content. This observation is in agreement with report by Douwes et al., (21) in which elevated numbers of tightly packed mitochondria were observed in SDHD linked head and neck PGLs. A similar increase in activities of Complex I, III and CS activities in SDH-related tumours when compared to VHL-related ones was

Figure 3 Relationship between tumour tissue catecholamine content and activity of respiratory chain enzyme complex II and tumour ATP/ADP/AMP content. Passing-Bablok regression plots for activities respiratory chain enzyme complex II (mU per mg protein) versus total catecholamine content (nmol per mg tissue) in PGL tumour tissues represented in Log10 scale. Points for samples below detection limit of 1H NMR spectroscopy are encircled in black.

33

Energy Metabolism and Catecholamine Content in PGL

2

observed by Fliedner et al., (22). All these factors suggest a compensatory response to the lack of SDH activity and associated activity of complex II. However as currently shown, the apparently increased activity of complex I, III, IV and CS in the tumours does not lead to full restoration of ATP/ADP/AMP. This contrasts with the report of Favier et al., (7)E in which no differences for complex III activity were observed among NF1/RET, SDH and VHL-related PGLs and that of Rapizzi et al., (10) where no differences in CS activity were observed across different genotypes. Our findings of higher complex IV activity in SDH-related tumours than NF1 and RET related ones contradicts with the findings by Favier et al., (7) who observed a decreased Complex IV activity in SDHx and VHL-related tumours. However, our observations are in agreement with the increased expression of complex IV protein reported in four out of five SDH-related tumours by Rapizzi et al., (10). Nevertheless, this study also indicated a high degree of variability in complex IV activities in these tumours. Previous studies by Lopez-Jimenez et al., (23) and Favier et al. (7) have demonstrated that there is a preferential activation of HIF-1α target genes such as glycolytic pathway enzymes in VHL tumours while activation of HIF-2α is observed in both VHL and SDHx tumours. It is also thought that activation of hypoxic response may be directly involved in decreased mitochondrial respiration. Interestingly, in the present study we observed an increase in complex I, III and IV and CS in SDHx tumours while there was a trend of overall down regulation of respiratory chain activities in VHL tumours. Also, we observed that in contrast to VHL tumours, accumulation of ATP/ADP/AMP in SDHx tumours was undetectable (except in one sample). This indirectly supports the observations by Favier et al., (7) that there is an activation of glycolysis as evidenced by increased lactate dehydrogenase activity in VHL tumours and not SDHx tumours. However, increased expressions of GLUT-1, GLUT-3 and Hexokinase II mRNAs observed in SDH tumours (7) can explain the high sensitivity of [18F]-FDG PET to SDH tumours (24-26). In in vitro experiments with cells silenced for SDH, it was observed that the accumulation of succinate leads to stabilization of HIF-1α (11). Later, Pollard et al., (27) described succinate accumulation in tumour tissues with germline SDH mutation, however, it was described in a single patient with pathogenic SDH mutation. In the present study, we for the first time provide strong in vivo evidence to support the hypothesis that there is an accumulation of succinate in SDH-related PGL tumour tissues. This supports the concept that stabilization of HIF-α in SDH- related tumours reflects the inhibitory effects of succinate on prolyl hydroxylase. High succinate accumulation observed in tumours with SDHx mutations also makes this a reliable parameter to indicate mutations in SDH subunits or assembly factors.

34

Genotype specific differences in catecholamine production by PGLs are well known. In the present study, our observations that tumours with mutations in VHL and SDHx mainly produce noradrenaline and those with mutations in RET and NF1 produce both adrenaline and noradrenaline support the previous findings by Eisenhofer et al.,(13). This difference in the catecholamine phenotype has been attributed to the lack of the enzyme phenylethanolamine N-methyltransferase (PNMT) in VHL and SDHx tumours (28). In line with the previous studies (13), we also observed that the tumour tissue total catecholamine content was lower in SDHx tumours when compared to RET and sporadic PGLs. These differences could be possibly attributed to the increased tyrosine hydroxylase activity in RET tumours (28) however, tyrosine hydroxylase activity or expression has not been investigated in SDHx tumours in comparison with other genotypes albeit a lack of the expression of this enzyme has been observed in biochemically silent SDHB tumours (29). In the present study we established that tumour tissue concentrations of ATP/ADP/AMP, as determined by low peak heights at relevant resonance positions in 1H-NMR spectra, are lower in SDH-related tumours than in other tumours. Further, our findings of positive relationships between respiratory enzyme complex II function, tumour ATP/ADP/AMP content and tumour catecholamine contents suggest the possibility that differences in energy metabolism might also contribute to the lower tumour tissue catecholamine contents in cluster 1 than in cluster 2 tumours. To this end, it is well known that the sequestration of catecholamines into secretory vesicles and re-uptake of catecholamines into chromaffin cells are active energy dependent processes. Sequestration of catecholamines is facilitated by vesicular monoamine transporters (VMATs). The H+ gradient necessary to maintain the activity of VMATs is generated by ATP dependent vesicular membrane proton pump (30). Chromaffin granules also contain strikingly high concentrations of ATP due to the activity of vesicular nucleotide transporter (31). This contributes to the stability and ability of chromaffin granules to maintain stores of catecholamines (32, 33). Further, norepinephrine transporter (NET) responsible for the sodium-chloride (Na+/Cl–)-dependent reuptake of extracellular norepinephrine and dopamine is also indirectly dependent on cellular store of ATP. NET functions by coupling the transport of norepinephrine and dopamine with the influx of sodium and chloride (Na+/Cl–). The ion gradients of Na+ and Cl- generated by the Na+/K+-ATPase make this reuptake energetically favorable (34, 35). Clearly therefore, energy metabolism has an important role in maintaining the stability of chromaffin granules and thus catecholamine storage. It thereby seems possible that genotype specific differences in the energy metabolism, along with associated differences in expression of various genes encoding components of secretory pathway and

35

Energy Metabolism and Catecholamine Content in PGL

2

exocytotic machinery (36), could in conjunction contribute to genotype specific differences in tumour catecholamine phenotypic features. In the present study we included 35 sporadic tumours 60% of which were not tested for SDHA and SDHAF-2. Two of the 3 sporadic tissues which had low respiratory chain complex II activities comparable to SDHx tumours also belong to the groups which were not tested for SDHA and SDHAF-2. Thus, mutations in SDHA and SDHAF-2 cannot be ruled out in these tumours. Also, the low activity may indicate that these tumours may have an as yet unidentified intronic or promotor mutations in SDH subunit genes or unidentified mutations in assembly factors genes. We used 1H NMR spectroscopy to determine the tumour tissue metabolite concentrations as it provides a holistic view on the tumour metabolome. This technique can very well identify various metabolites and quantify differences in the metabolite concentrations among different samples, but it is limited by its sensitivity. It can quantify metabolites only in micromolar range because of which, many intermediates of energy and catecholamine metabolism could not be determined in the present study. This is clearly visible in the Passing-Bablok regression analysis of total catecholamines vs Complex II activity and ATP/ADP/AMP levels where reduced sensitivity of NMR spectroscopy separates out a group of samples which if analyzed with a more sensitive method could have reflected the linear relationship better. Nevertheless, the study was successful in identifying the relationship between catecholamine content of PGLs and energy metabolism. Catecholamine contents are particularly low in tumours due to SDHB mutations and it has been suggested that this along with diversion of energy from maintaining catecholamine phenotypic features to growth might contribute to the larger sizes and more aggressive features of these tumours (18). The present study, establishing relationships between tumour energetics and catecholamine phenotypic features, provides new insight into how such diversions of energy might occur with implications for novel therapeutic strategies targeting energy pathways.

36

2.5 References1. Lenders JW, Eisenhofer G, Mannelli M, Pacak K. Phaeochromocytoma. Lancet 2005;366: 665-75.2. Welander J, Soderkvist P, Gimm O. Genetics and clinical characteristics of hereditary pheochromocytomas

and paragangliomas. Endocrine-related cancer 2011;18: R253-76.3. Burnichon N, Buffet A, Parfait B, Letouze E, Laurendeau I, Loriot C, et al. Somatic NF1 inactivation is a

frequent event in sporadic pheochromocytoma. Human molecular genetics 2012; 21:5397-405.4. Burnichon N, Vescovo L, Amar L, Libe R, de Reynies A, Venisse A, et al. Integrative genomic analysis

reveals somatic mutations in pheochromocytoma and paraganglioma. Human molecular genetics 2012;20: 3974-85.

5. Zhuang Z, Yang C, Lorenzo F, Merino M, Fojo T, Kebebew E, et al. Somatic HIF2A gain-of-function mutations in paraganglioma with polycythemia. The New England journal of medicine 2012;367: 922-30.

6. Dahia PL. Transcription association of VHL and SDH mutations link hypoxia and oxidoreductase signals in pheochromocytomas. Annals of the New York Academy of Sciences 2006;1073: 208-20.

7. Favier J, Briere JJ, Burnichon N, Riviere J, Vescovo L, Benit P, et al. The Warburg effect is genetically determined in inherited pheochromocytomas. PloS one 2009;4: e7094.

8. Gimenez-Roqueplo AP, Favier J, Rustin P, Mourad JJ, Plouin PF, Corvol P, et al. The R22X mutation of the SDHD gene in hereditary paraganglioma abolishes the enzymatic activity of complex II in the mitochondrial respiratory chain and activates the hypoxia pathway. American journal of human genetics 2001;69: 1186-97.

9. Gimenez-Roqueplo AP, Favier J, Rustin P, Rieubland C, Kerlan V, Plouin PF, et al. Functional consequences of a SDHB gene mutation in an apparently sporadic pheochromocytoma. The Journal of clinical endocrinology and metabolism 2002;87: 4771-4.

10. Rapizzi E, Ercolino T, Canu L, Giache V, Francalanci M, Pratesi C, et al. Mitochondrial function and content in pheochromocytoma/paraganglioma of succinate dehydrogenase mutation carriers. Endocrine-related cancer 2012;19: 261-9.

11. Selak MA, Armour SM, MacKenzie ED, Boulahbel H, Watson DG, Mansfield KD, et al. Succinate links TCA cycle dysfunction to oncogenesis by inhibiting HIF-alpha prolyl hydroxylase. Cancer cell 2005;7: 77-85.

12. Lee S, Nakamura E, Yang H, Wei W, Linggi MS, Sajan MP, et al. Neuronal apoptosis linked to EglN3 prolyl hydroxylase and familial pheochromocytoma genes: developmental culling and cancer. Cancer cell 2005;8: 155-67.

13. Eisenhofer G, Pacak K, Huynh TT, Qin N, Bratslavsky G, Linehan WM, et al. Catecholamine metabolomic and secretory phenotypes in phaeochromocytoma. Endocrine-related cancer;18: 97-111.

14. Rodenburg RJ. Biochemical diagnosis of mitochondrial disorders. Journal of inherited metabolic disease 2010;34: 283-92.

15. Watanabe N, Kamei S, Ohkubo A, Yamanaka M, Ohsawa S, Makino K, et al. Urinary protein as measured with a pyrogallol red-molybdate complex, manually and in a Hitachi 726 automated analyzer. Clinical chemistry 1986;32: 1551-4.

16. Engelke UF, Kremer B, Kluijtmans LA, van der Graaf M, Morava E, Loupatty FJ, et al. NMR spectroscopic studies on the late onset form of 3-methylglutaconic aciduria type I and other defects in leucine metabolism. NMR in biomedicine 2006;19: 271-8.

17. Wevers RA, Engelke U, Wendel U, de Jong JG, Gabreels FJ, Heerschap A. Standardized method for high-resolution 1H-NMR of cerebrospinal fluid. Clinical chemistry 1995;41: 744-51.

18. Eisenhofer G, Lenders JW, Siegert G, Bornstein SR, Friberg P, Milosevic D, et al. Plasma methoxy-tyramine: a novel biomarker of metastatic pheochromocytoma and paraganglioma in relation to established risk factors of tumour size, location and SDHB mutation status. Eur J Cancer 2012;48: 1739-49.

19. Eisenhofer G, Goldstein DS, Stull R, Keiser HR, Sunderland T, Murphy DL, et al. Simultaneous liquid- chromatographic determination of 3,4-dihydroxyphenylglycol, catecholamines, and 3,4-dihydroxy-phenylalanine in plasma, and their responses to inhibition of monoamine oxidase. Clinical chemistry 1986;32: 2030-3.

37

Energy Metabolism and Catecholamine Content in PGL

2

20. Passing H, Bablok. A new biometrical procedure for testing the equality of measurements from two different analytical methods. Application of linear regression procedures for method comparison studies in clinical chemistry, Part I. Journal of clinical chemistry and clinical biochemistry 1983;21: 709-20.

21. Douwes Dekker PB, Hogendoorn PC, Kuipers-Dijkshoorn N, Prins FA, van Duinen SG, Taschner PE, et al. SDHD mutations in head and neck paragangliomas result in destabilization of complex II in the mitochondrial respiratory chain with loss of enzymatic activity and abnormal mitochondrial morphology. The Journal of pathology 2003;201: 480-6.

22. Fliedner SM, Kaludercic N, Jiang XS, Hansikova H, Hajkova Z, Sladkova J, et al. Warburg effect’s manifestation in aggressive pheochromocytomas and paragangliomas: insights from a mouse cell model applied to human tumour tissue. PloS one 2012;7: e40949.

23. Lopez-Jimenez E, Gomez-Lopez G, Leandro-Garcia LJ, Munoz I, Schiavi F, Montero-Conde C, et al. Research resource: Transcriptional profiling reveals different pseudohypoxic signatures in SDHB and VHL-related pheochromocytomas. Molecular endocrinology 2010; 24(12):2382-91.

24. Timmers HJ, Chen CC, Carrasquillo JA, Whatley M, Ling A, Eisenhofer G, et al. Staging and functional characterization of pheochromocytoma and paraganglioma by 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography. Journal of the National Cancer Institute 2012;104: 700-8.

25. Timmers HJ, Chen CC, Carrasquillo JA, Whatley M, Ling A, Havekes B, et al. Comparison of 18F-fluoro- L-DOPA, 18F-fluoro-deoxyglucose, and 18F-fluorodopamine PET and 123I-MIBG scintigraphy in the localization of pheochromocytoma and paraganglioma. The Journal of clinical endocrinology and metabolism 2009;94: 4757-67.

26. Timmers HJ, Kozupa A, Chen CC, Carrasquillo JA, Ling A, Eisenhofer G, et al. Superiority of fluorodeoxy-glucose positron emission tomography to other functional imaging techniques in the evaluation of metastatic SDHB-associated pheochromocytoma and paraganglioma. J Clin Oncol 2007;25: 2262-9.

27. Pollard PJ, Briere JJ, Alam NA, Barwell J, Barclay E, Wortham NC, et al. Accumulation of Krebs cycle intermediates and over-expression of HIF1alpha in tumours which result from germline FH and SDH mutations. Human molecular genetics 2005;14: 2231-9.

28. Eisenhofer G, Walther MM, Huynh TT, Li ST, Bornstein SR, Vortmeyer A, et al. Pheochromocytomas in von Hippel-Lindau syndrome and multiple endocrine neoplasia type 2 display distinct biochemical and clinical phenotypes. The Journal of clinical endocrinology and metabolism 2001;86: 1999-2008.

29. Timmers HJ, Pacak K, Huynh TT, Abu-Asab M, Tsokos M, Merino MJ, et al. Biochemically silent abdominal paragangliomas in patients with mutations in the succinate dehydrogenase subunit B gene. The Journal of clinical endocrinology and metabolism 2008;93: 4826-32.

30. Henry JP, Sagne C, Bedet C, Gasnier B. The vesicular monoamine transporter: from chromaffin granule to brain. Neurochemistry international 1998;32: 227-46.

31. Sawada K, Echigo N, Juge N, Miyaji T, Otsuka M, Omote H, et al. Identification of a vesicular nucleotide transporter. Proceedings of the National Academy of Sciences of the United States of America 2008;105: 5683-6.

32. Winkler H. The composition of adrenal chromaffin granules: an assessment of controversial results. Neuroscience 1976;1: 65-80.

33. Kopell WN, Westhead EW. Osmotic pressures of solutions of ATP and catecholamines relating to storage in chromaffin granules. The Journal of biological chemistry 1982;257: 5707-10.

34. Galli A, DeFelice LJ, Duke BJ, Moore KR, Blakely RD. Sodium-dependent norepinephrine-induced currents in norepinephrine-transporter-transfected HEK-293 cells blocked by cocaine and antide-pressants. The Journal of experimental biology 1995;198: 2197-212.

35. Torres GE, Gainetdinov RR, Caron MG. Plasma membrane monoamine transporters: structure, regulation and function. Nature reviews 2003;4: 13-25.

36. Eisenhofer G, Huynh TT, Elkahloun A, Morris JC, Bratslavsky G, Linehan WM, et al. Differential expression of the regulated catecholamine secretory pathway in different hereditary forms of pheochromocytoma. American journal of physiology 2008;295: E1223-33.

38

Supplementary Figure 1 Examples of 1H NMR spectra for different genotypes A. Example of NMR spectrum of a sporadic tumour depicting the peak positions of lactic acid, acetic acid and catecholamines. B. Example of NMR spectrum of a SDHB tumour depicting peak position of succinic acid. C. Example of NMR spectrum of a RET tumour depicting ATP/ADP/AMP and adrenaline specific peaks.

A

B

C

39

Energy Metabolism and Catecholamine Content in PGL

2

Supplementary Figure 2 Comparison of epinephrine and norepinephrine content in tumour tissues as determined by HPLC and 1H-NMR spectroscopy. Passing - Bablok regression plot to compare measurement of epinephrine and norepinephrine content in PGL tumour tissues by HPLC and 1H-NMR spectroscopy expressed as µg per mg tissue in Log10 scale.

Supplementary Figure 3 Activities of respiratory chain complex II and tumour tissue succinate content in SDH related tumours. Black bars represent complex II activity expressed in mU/mg protein and grey bars represent tumour tissue succinate content in mmol/mg tissue for each individual tumour. Represented on X axis is the hereditary mutation in the patient from whom the tumour was resected.

Genotype–specific differences in the tumour metabolite profile of pheochromocytoma and paraganglioma using untargeted and targeted metabolomics

Rao JU, Engelke UFH, Sweep FCGJ, Pacak K, Kusters B, Goudswaard AG, Hermus ARMM, Mensenkamp AR, Eisenhofer G, Qin N, Richter S, Kunst HPM, Timmers HJLM, Wevers RA.

J Clin Endocrinol Metab. 2015 100(2):E214-22.

3

42

Abstract

Context and Objective: Pheochromocytomas and paragangliomas (PGLs) are neuroendocrine tumours of sympathetic or parasympathetic paraganglia. Nearly 40% of PGLs are caused by germline mutations. The present study investigated the impact of genetic alterations on metabolic networks in PGLs. Design: Homogenates of 32 sporadic PGLs and 48 PGLs from patients with mutations in SDHB, SDHD, SDHAF-2, VHL, RET and NF-1 were subjected to 1H-NMR spectroscopy at 500 MHz for untargeted and HPLC tandem mass spectrometry for targeted metabolite profiling. Results: 1H-NMR spectroscopy identified 28 metabolites in PGLs of which 12 showed genotype-specific differences. Part of these results published earlier reported low complex II activity (p<0.0001) and low ATP/ADP/AMP content (p<0.001) in SDH-related PGLs compared to sporadics and PGLs of other genotypes. Extending these results, low levels of N-acetylaspartic acid (NAA, p<0.05) in SDH tumours and creatine (p<0.05) in VHL tumours were observed compared to sporadics and other genotypes. Positive correlation was observed between NAA and ATP/ADP/AMP content (p<0.001) and NAA and complex II activity (p<0.0001) of PGLs. Targeted purine analysis in PPGLs showed low adenine in cluster 1 compared to cluster 2 tumours (SDH p<0.0001; VHL p<0.05) while lower levels (p<0.05) of guanosine and hypoxanthine were observed in RET tumours compared to SDH tumours. Principal component analysis (PCA) of metabolites, could distinguish PGLs of different genotypes. Conclusions: The present study gives a comprehensive picture of alterations in energy metabolism in SDH- and VHL-related PGLs and establishes the interrela-tionship of energy metabolism and amino acid and purine metabolism in PGLs.

43

Genotype Specific Differences in PGL Metabolome

3

3.1 Introduction

Pheochromocytomas and paragangliomas (PGLs), are neuroendocrine tumours of the adrenal medulla and sympathetic or parasympathetic paraganglia (1). Nearly 40% of them are caused by germline mutations in tumour susceptibility genes which include VHL (von Hippel-Lindau), RET (rearranged during transfection), NF-1 (neurofibromatosis type 1), SDHA/B/C/D/AF2 (succinate dehydrogenase subunits A, B, C and D and assembly factor 2), TMEM127 (transmembrane protein 127), MAX (myc-associated factor X) (2) and more recently discovered HIF2α (Hypoxia inducible factor 2 α) (3) and FH (fumarate hydratase) (4). Gene expression profiling has revealed the presence of two clusters of PGLs: cluster 1 (VHL, SDH) and cluster 2 (RET, NF-1, TMEM127 and MAX) displaying signature of transcripts associated with angiogenesis and hypoxia and RNA synthesis and kinase signaling, respectively. Previous studies by us and others demonstrate a genotype-specific alteration in respiratory chain enzyme activities (5-7). In comparison with other genotypes, VHL tumours show an overall decrease in respiratory chain enzyme function, while SDH tumours demonstrate a decreased complex II and an attempted compensatory increase in the complex I, III and IV activities, indicative of impaired mitochondrial function (6). Mitochondria are the metabolic hubs of the tumour which in addition to providing energy, generate reducing equivalents (NADH and NADPH) and intermediates for anabolic reactions. Thus, genotype-specific differences in the mitochondrial function could have a strong impact on inter- mediary metabolism of PGLs beyond aerobic glycolysis. In this scenario, use of metabolomic techniques help to discover and systematically analyze the metabolic fingerprint of the tumours thus giving insights into alterations in the metabolic pathways that are involved in tumourigenesis (8). Proton (1H) nuclear magnetic resonance (NMR) spectroscopy is a cornerstone technology for untargeted metabolomics. It provides a holistic view of metabolism and has other advantages like relatively simple sample preparation and rapid data acquisition. Moreover, it allows quantitation of large number of metabolites using a single internal standard (9,10). In the present study, we used a combination of 1H NMR spectroscopy and liquid chromatography (LC) tandem mass spectrometry (MS) to investigate geno-type-specific differences in the metabolic profile of 80 PGLs. 1H NMR spectroscopy was used for untargeted metabolite profiling and LC tandem MS was used to further investigate and/or validate the findings obtained from 1H NMR spectroscopy.

44

3.2 Materials and Methods

Patients In the present study, tumours of patients with histologically proven PGLs evaluated at the Radboud University Medical Centre (RUMC), Nijmegen, the Netherlands (n=52) and Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) (n=24), National Institute of Health (NIH), Bethesda, MD, USA who underwent surgical resection between 1988 and 2012 and between 2003 and 2010 respectively, were included. Frozen primary tumours were used for the study. No metastatic tumours were used. The presence of germline mutations and large deletions in SDHB/C/D, RET, VHL and -since 2011- in SDHA, SDHAF2, TMEM127 and MAX, was investigated using standard procedures. Data were collected under conditions of regular clinical care, with ethical committee approval for the use of those data for scientific purposes at RUMC. The study was approved by the Institutional Review Board of NICHD, and all patients gave written informed consent before testing. The details of the patients’ clinical characteristics, tumour location and genotype are listed in Table 1. Tumours from patients in whom germline mutation could not be found in any of the PGL-genes were considered “sporadic”. The extraadrenal tumours were mainly abdominal (retroperitoneal) or thoracic in location.

Tumour tissue processing and sample preparationTumours were procured as early as possible, dimensions were recorded and a small piece of the tumour was weighed, snap frozen, stored in liquid nitrogen and used for experimental purposes. For histological confirmation, additional slices were stained with hematoxylin and eosin and re-evaluated by an independent pathologist (BK). The tumour tissues were homogenized on ice in 10µl of ice cold distilled water per mg tissue using a hand-held Teflon/glass homogenizer. The samples were then centrifuged at 16,000xg for 10min at 4oC and the supernatants were subjected to ultrafiltration with Vivaspin Turbo15, 10kDa filters (Sartorius, Germany) for deproteinization. 1H NMR Spectroscopy One-dimensional 1H NMR spectroscopy was performed to investigate the metabolic profile of PGLs. For this purpose, the total ultrafiltrate was made up to 700µl with water, pH was adjusted to 2.5 and 20µl of 20.2mM sodium 3-trimethyl-silyl-2,2,3,3-tetradeuteropropionate (TSP, Aldrich, Zevenaar, The Netherlands) in D2O (Catalogue No. 435767, Aldrich, Zevenaar, The Netherlands) was added to the samples. The samples were then placed in 5mm NMR tubes and 1H NMR spectra

45

Genotype Specific Differences in PGL Metabolome

3

were obtained using a Bruker 500MHz spectrometer (pulse angle 90o, delay time 4s, number of scans 256). Water resonance was suppressed by gated irradiation centred on the water frequency. For resonance assignment, 2D COSY NMR spectra were acquired for some of the samples. The spectral width in the F1 and F2 domains were 5500Hz. A total of 2K data points were collected in t2, 256 t1 increments with 32 transient per increment were used. The relaxation delay was set to 2s. Before the Fourier transformation, a sine-bell function was applied in both time domains. During the relaxation delay, the water resonance was presaturated. Analysis of NMR spectraThe free induction decays (FIDs) measured for these samples were processed using Topspin software (Bruker, Germany). Fourier transformation was applied on the FID of the samples and the resulting spectra were phase and baseline corrected. The chemical shifts in the spectra were referenced to the internal standard, TSP. Assignment of peak positions for compound identification was performed by comparing the peak positions in the spectra of PGLs with the reference spectral database of model compounds at pH 2.5 using Amix version 3.9.14 (BRUKER BIOSPIN GmbH, Rheinstetten, Germany) (11,12). Quantification of identified compounds was performed by manual integration of chosen peak/s for a specific metabolite. Further confirmation of metabolite identification based on peak positions was obtained from 2D NMR.

Table 1 Patient Characteristics

Genotype N (patients)

Age y(mean±SD)

Gender (M/F)

N (tumours)

Tumour location (A/E/HN)

Tumour volume (cm3)

Sporadic 32 45.8 ± 13.9 15/17 32 29/3/0 227.7 ± 480.1

SDHB 9 31.6 ± 11.2 8/1 11 1/10/0 295.7 ± 335.5

SDHD 8 44 ± 10.5 6/2 8 1/1/6 27.9 ± 47.2

SDHAF2 1 24 1/0 1 0/0/1 18.9

VHL 7 31.4 ± 10.5 6/1 7 6/1/0 70.9 ± 97.8

RET 12 37.7 ± 13.3 6/6 14 14/0/0 97.6 ± 179.4

NF1 7 43.1 ± 18.9 4/3 7 7/0/0 115.7 ± 85.2

N: Number, y: years, M: Male, F: Female, A: Adrenal, E: Extra-adrenal, HN: head and neck

46

High pressure liquid chromatography (HPLC) tandem mass spectroscopy (MS):Deproteinized tumour tissue lysate was used for targeted measurement of purines using LC tandem mass spectrometry. Only PGLs with known hereditary mutations (n=47, one VHL tumour excluded for limited sample quantity) were included in this experiment. The samples were prepared by adding 100µl of stable isotope internal standard solution containing dihydrouracil, uracil, uric acid, orotic acid, dihydrothymine, thymine, guanosine and thymidine and 10µl of 2% formic acid for acidification to 100µl of deproteinized tumour tissue lysate. 5µl of the sample thus prepared was injected into the HPLC tandem MS system. The HPLC tandem MS process is described in detail in supplementary information.

Multivariate Statistical AnalysisA total of 45 1H NMR spectra corresponding to 13 RET, 7 NF-1, 18 SDH and 7 VHL tumours were considered for PCA. Spectra from 1 RET and 2 SDH tumours were omitted in the analysis as the quality of the spectra was not suitable for PCA. PCA was also performed on purine concentrations normalized around mean from 20 SDH, 6 VHL, 14 RET and 7 NF-1. For details see supplementary methods.

Confirmation of metabolites identified by 1H-NMR spectroscopyIndependent targeted and quantitative HPLC and LC tandem MS methods were used for confirmation of detection and quantitation of succinate and N-acetylas-partic acid (NAA) (see supplementary information).

Statistical methodsStatistical analyses were performed using SPSS (SPSS Inc., v.18, Chicago, IL) and GraphPad Prism 6 software (GraphPad, La Jolla, CA). The data was analyzed using independent samples Kruskal-Wallis test and Dunn’s post-test was used to compare the different genotypes. Statistical significance was accepted at P <0.05.

3.3 Results

Overview of 1H-NMR spectra of PGLsNMR spectroscopy was performed in all 80 PGLs. The spectra showed resonances of 28 different metabolites (Fig 1 and Table 2). Among these are amino acids, sugars, intermediates of Krebs cycle and glycolysis, catecholamines, organic acids, high energy phosphates, creatine/creatinine, NAA and N-acetylaspartylglutamic acid.

47

Genotype Specific Differences in PGL Metabolome

3

Genotype-specific differences in the metabolites as identified by 1H-NMR spectroscopyNAA content of the tumour tissue was estimated by integrating the area under the peak at 2.03ppm. This peak was chosen as the multiplets at 2.94 and 4.72 overlapped with those of N-acetylaspartylglutamic acid. NAA was below detection limit in all the SDH and VHL tumours tested except one (Fig 2E, supplementary fig 4). The levels were higher in RET (p<0.01), NF-1 (p<0.05) and sporadic tumours (p<0.05) when compared to SDH tumours (Fig 2E). NAA peaks were undetectable in 16 sporadics, 3 RET and 3 NF-1 tumours. HPLC tandem MS was used to validate the NAA content of the tumours and a significant correlation and linear relationship was observed between the two methods (supplementary fig 1). Creatine concentrations in tumours were estimated by integrating the area under peak at 3.05ppm as the singlet at 4.11ppm had a lower signal to noise ratio and also

Figure 1 One dimensional 500 MHz 1H NMR spectra of PGLs tumour tissues at pH 2.50 A. Sporadic tumour showing normal tumour metabolites such as lactic acid, acetic acid and catecholamines. B. SDHB tumour showing a high concentration of succinic acid (9.89 nmol/mg tissue). C. RET tumour showing high resonance of epinephrine and the overlap of the three doublets deriving from ATP/ADP/AMP.

A

B

C

48

Table 2 NMR assignmenta (pH 2.50) and occurrence for metabolites observed in PGLs of different genotypes

Metabolites δ 1H Genotype

SDH VHL RET NF-1 SporadicAcetic acid 2.08 s ++ ++ ++ ++ ++

N- Acetylaspartic acid 2.03 s - + + + +N- Acetylglutamylasparagine 2.05 s - - + (n=6) - + (n=4)4- Aminobutyric acid 2.50 t + + + + +

ATP/ADP/AMP 6.20b d - + + + +Citric acid 2.92 AB - - - - + (n=4)Creatine 3.05 s; 4.11 s ++ + ++ ++ ++Creatinine 3.13 s; 4.29 s ++ + ++ ++ ++Epinephrine 2.73 t - - ++ ++ ++/-Ethanol 1.17 t; 3.64 q ++ ++ ++ ++ ++Formic acid 8.24 s + + + + +Glutamate/Glutamine 2.17 m + + + + +Histidine 8.67 d - + (n=1) + - -

3- Hydroxybutyric acid 1.23 d + + + + +Inosine 6.09 d; 8.22 s + ++ ++ ++ ++Inositol-myo 3.27 t; 4.05 t + + + + +Inositol-scyllo 3.34 s + + + + +Isoleucine 0.94 t; 1.02 d + + - + +Lactic acid 1.41d; 4.36 q +++ +++ +++ +++ +++Lysine 1.72 m + + + + +Methanol 3.35 s ++ ++ ++ ++ ++Methionine 2.12 s + + + + +Norepinephrine 6.90 m + + ++ ++ ++Pyruvate 2. 37 s + + + + +Succinic acid 2.67 s +++ + + + +Sucrose 5.40 d - - - - ++ (n=2)Threonine 1.33 d + + + + +Unknown 2.14 s + - - - -Valine 1.00 d; 1.04 d + + + + +

+: Present in spectrum, +: low; ++: medium +++: high; based on peak heights– : Absent in spectrum

a: Chemical shifts (δ) in ppm relative to TSP; multiplicity are s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet; AB, AB system (second order coupling).

b: overlap of three doubletsn: Number of patients in whom the metabolites was found

49

Genotype Specific Differences in PGL Metabolome

3

Figu

re 2

Met

abol

ite p

rofil

es a

nd P

CA a

naly

sis o

f 1 H N

MR

spec

tra

of P

GLs

bel

ongi

ng to

dif

fere

nt g

enot

ypes

. Dot

plo

t rep

rese

ntin

g A

. N-a

cety

l-as

part

ate

B. c

reat

ine

and

C. is

oleu

cine

in m

g pe

r mg

tum

our t

issu

e. H

oriz

onta

l lin

e re

pres

ents

the

mea

n. D

ata

sets

hav

ing

diff

eren

t alp

habe

ts

abov

e th

em a

re s

igni

fican

tly

diff

eren

t wit

h da

ta s

ets

havi

ng th

e al

phab

et ‘a

’ bei

ng th

e lo

wes

t (p<

0.05

). In

40

tum

our s

ampl

es N

-ace

tyla

spar

-ta

te, i

n 4

tum

our s

ampl

es c

reat

ine

and

in 3

2 tu

mou

r sam

ples

isol

euci

ne w

ere

belo

w d

etec

tion

lim

it. T

hey

have

bee

n re

pres

ente

d as

hal

f the

lo

wes

t det

ecta

ble

valu

e. T

he P

CA sc

ore

plot

, D, r

epre

sent

s the

firs

t and

seco

nd p

rinc

iple

com

pone

nts a

nd c

lear

ly d

iscr

imin

ate

betw

een

SDH

, RET

an

d V

HL t

umou

rs. T

he lo

adin

g pl

ot, E

, sho

ws t

he p

eak

posi

tion

s con

trib

utin

g to

the

max

imum

var

ianc

es in

pri

ncip

al co

mpo

nent

s 1 a

nd 2

in th

e ou

ter c

ircl

e in

dica

ting

that

the

reso

nanc

es o

f suc

cini

c ac

id, A

TP/A

DP/

AM

P, ep

inep

hrin

e an

d th

e un

know

n si

ngle

t res

onan

ce a

t 2.14

ppm

cau

se

the

sepa

rati

on b

etw

een

RET

and

SDH

tum

ours

in th

e PC

A sc

ore

plot

.

AB

DE

C

50

was in a region with peaks arising from several other metabolites. Creatine levels were high in NF-1 tumours when compared to SDHD/AF2 head and neck PGLs (p<0.01), VHL (p=0.0001) and RET (p<0.05) tumours (supplementary fig 4). VHL tumours had low creatine levels when compared to sporadic tumours (p<0.05) (Fig 2F). Creatine peaks were undetectable in 1 sporadic and 2 RET related PGLs. Isoleucine levels in the tumour tissue were determined by integrating area under peaks for the doublet at 1.02ppm as this doublet resonance had the best signal to noise ratio. Isoleucine peaks were consistently undetectable in RET tumours while in the tumours of other genotypes the levels were measurable (Fig 2D) though its levels were found to be undetectable in 9 sporadic, 5 SDH, 3 VHL and 3 NF-1 tumours. The as yet unassigned singlet resonance at 2.14ppm occurs in all SDH tumours. The concentration of the compound causing this resonance must be in the low micromolar range. As the singlet partially co-resonates with the glutamine multiplet resonance at 2.17ppm it is hard to decide whether the 2.14ppm singlet also is present in lower concentration in the other tumour types. Clearly, the concentration of the metabolite causing the 2.14ppm singlet in the SDH tumours is significantly higher than in the other tumour types.

Correlation of NAA content with energy metabolism in PGLsPositive correlations were observed between NAA and ATP/ADP/AMP content (R=0.422, p<0.001) and NAA content and activity of respiratory chain complex II (R=0.465, p<0.0001) in PGL tumours.

Multivariate statistical analysis of NMR resultsUsing PCA without scaling, tumours with mutations in SDHB, VHL and RET could be well separated based on the variances in the peak positions pertaining to epinephrine, succinate, ATP/ADP/AMP, an unknown peak at 2.14ppm and lactic acid (Figure 2D and E). High levels of epinephrine in RET tumours, increased succinate accumulation and low levels of ATP/ADP/AMP in SDH tumours and higher levels of lactic acid in VHL tumours when compared to RET tumours are identified in the PCA as most discriminating features. NF-1 tumours from two patients were classified with VHL tumours as they contained higher lactic acid levels and epinephrine levels below the detection limit of NMR spectroscopy. The NF-1 tumours did not form a separate group as there was no single distinguishing metabolite detectable by 1H-NMR spectroscopy for this group. To highlight the subtle differences and to minimize the effect of high intensity peaks, PCA was also performed after autoscaling which revealed threonine (1.33ppm) and alanine (1.51ppm) peaks contributing to the separation of SDH tumours from the RET tumours.

51

Genotype Specific Differences in PGL Metabolome

3

Genotype-specific differences in Purine metabolism in PGLsObservations of consistent lack of ATP/ADP/AMP peaks in SDH tumours when compared to the other genotypes could probably be related to altered purine biosynthetic pathways in SDH related PGLs. Thus, purine concentrations were estimated using HPLC-tandem MS in PGL lysates prepared for 1H NMR spectroscopy. Adenine levels were lower in SDH tumours when compared to RET and NF-1 tumours (p<0.0001 and p<0.05, respectively). Adenosine levels were lower in SDH (p<0.01) and VHL (p<0.05) tumours when compared to RET tumours in contrast to guanosine which was higher in SDH tumours when compared to RET tumours (p<0.05). Inosine levels were lower in cluster 1 tumours compared to those belonging to cluster 2 (p<0.01) differing from hypoxanthine which was higher in SDH tumours when compared to RET related ones (p<0.05, Table 3).

Multivariate statistical analysis of purinesUsing PCA, the samples with mutations in SDH, RET and NF-1 could be separated based on the differences in the concentrations of adenine, hypoxanthine and inosine (supplementary figure 2). High levels of adenine in RET and hypoxanthine in SDH tumours emerge as the most discriminating features for these tumours in the PCA analysis.

Table 3 Overview of purines (mg/mg tissue) detected in PGLs

Metabolites Genotype

Cluster 1 Cluster 2

SDH (n=20) VHL (n=6) RET (n=14) NF-1 (n=7)

Adenine 0.68 ± 0.80a 3.3 ± 4.11 a,c 16.18 ± 5.41b 8.33 ± 6.95b,c

Adenosine 1.33 ± 2.84 a 6.53 ± 12.67 a,b 3.59 ± 2.93b 5.65 ± 5.67 b

Guanine 7.3 ± 6.73 2.9 ± 3.29 3.08 ± 1.29 2.01 ± 1.10

Guanosine 4.44 ± 5.26 b 4.51 ± 5.54 a,b 0.86 ± 0.9a 1.79 ± 1.60 a,b

Inosine 23.35 ± 20.15 a 47.42 ± 36.10 a,b 40.26 ± 19.44a,b 78.32 ± 67.70 b

Hypoxanthine 103.73 ± 48.28 b 116.16 ± 50.70 a,b 64.16 ± 19.46 a 96.49 ± 47.22 a,b

Xanthine 14.52 ± 15.58 20.22 ± 15.47 12.71 ± 6.60 15.96 ± 6.87

Uric Acid 64.25 ± 147.32 11.58 ± 4.21 31.76 ± 63.92 23.10 ± 28.38

a and b: a Values are significantly lower than b values (where b is P< .05) when indicated after unique values. c No significant differences were observed between the two groups.

52

3.4 Discussion

The present study extends the results of genotype-specific differences in metabolic profiles of PGLs from our previous publication using the same set of tumours (6). In the previous study using the same set of PGLs we reported decrease in complex II activity (p<0.0001), increase in succinate content (p<0.001) and decrease in ATP/ADP/AMP content (p<0.001) in SDH tumours when compared to sporadic PGLs and PGLs of other genotypes. Further, we discussed the interrelationship of energy and catecholamine metabolism (6). In the present study, we have used multiple complementary analytical methods, to further investigate genotype-specific differences in tumour metabolite profile. Here we describe genotype-specific differences in the contents of additional metabolites and demonstrate that these differences in metabolite profiles can be used to differentiate the genotypes using unsupervised statistical methods. The study thus gives a comprehensive picture of alterations in energy metabolism in SDH and VHL tumours. Further, interrela-tionships of genotype-specific alterations in energy, amino acid and purine metabolism are established. Metabolite profiling of tumours has been approached using two basic technological platforms, NMR or MS. In the present study, we have used multiple complementary analytical techniques which use both of these platforms. We used NMR as an untargeted metabolite profiling technique allowing detection of the vast majority of proton containing compounds in the low micromolar or higher concentration range. For targeted metabolite profiling, we used HPLC tandem MS to test the presence of genotype-specific differences in the concentration of various purines in the tumours. NMR based metabolite profiling has the inherent disadvantage of limited low micromolar sensitivity which limits the number of metabolites that could be identified. Thus, we were unable to generate a metabolomic profile of sufficient complexity for unbiased application of pathway analysis tools to identify geno-type-specific dysfunctional metabolic pathways. We used PCA as an unsupervised method for interpretation of NMR spectra and to classify the samples in genotype-specific groups on the basis of their metabolite profile. The peak positions contributing to maximum variability in the first three principal components included epinephrine, succinate, lactic acid and high energy phosphates of adenosine. This is explained by their respective roles in catecholamine and energy metabolism. Interestingly, high lactate levels observed in VHL tumours support the observations of high LDH activity in these tumours by Favier et al (5). However, the results need to be interpreted with caution as lactate levels in resected samples do not necessarily reflect the actual tumour metabolism. Despite alanine and threonine emerging as important variables

53

Genotype Specific Differences in PGL Metabolome

3

after autoscaling, significant differences were not observed for these amino acids among the tumours of different genotypes when the concentrations of these amino acids were calculated based on the area under peak of the NMR spectra. Reduced tumour ATP/ADP/AMP content as measured by 1H-NMR spectroscopy, suggested that there could be a reduced total adenylate, adenosine and adenine content in the SDH tumours. This hypothesis was verified by assessing content of purines in PGLs using HPLC tandem MS. Based on the results of this study, we hypothesize that in the de novo synthesis of the purine nucleotides, possibly the activity of inosine monophosphate (IMP) dehydrogenase is higher in SDH tumours while in the RET related ones, adenylosuccinate synthase and/or lyase activities are higher. Thus, the IMP that is formed by the de novo pathway is committed to either of the purine biosynthetic branches in a genotype-specific way (Figure 3). In a similar situation, Pillwein et al., (13) compared the normal brain tissue and glioblastomas for pool of purines and purine nucleotides and nucleosides and the activities of IMP dehydrogenase, hypoxanthine guanine phosphoribosyl transferase (HGPRT) and other enzymes. It was observed that total adenylates and guanylates were lower by 60% and 71% respectively and that incubation of such tumour sections with Tiazofurin, an IMP dehydrogenase inhibitor lead to a reduction in total guanylate pool. Additionally, the enzyme HGPRT in the salvage pathway might have genotype-specific differences in substrate availability which leads to the differences in their purine nucleotide profile. A recent report by Gottle et al. (14) showed that increased purine nucleotide pools and energy reserve are observed post differentiation in rat pheochromocytoma cells (PC-12 subclones) which indirectly suggests that the lower purine pools are reflective of dedifferen-tiation. However, in contrast to the observations of Gottle et al., (14) that undiffer-entiated cells have lower dopamine, SDH tumours have high dopamine levels. Thus, it is not clear whether purine pool has an effect on dopamine levels in hereditary PGLs. Further, using PCA of purine levels in PGLs, SDH, RET and NF-1 tumours could be distinguished well. However, genotype-specific differences were not observed in levels of more downstream metabolites of purine biosynthetic pathway like xanthine and uric acid indicating lack of genotype-spe-cific differences in further purine catabolism. In this study, for the first time we report the presence of NAA in the peripheral nervous system. NAA is synthesized through acetylation of aspartate by L-Asparatate N-acetyl transferase, a membrane bound enzyme, in the neuronal tissue (16). NAA is synthesized and exported from mitochondria and its synthesis is dependent on mitochondrial energy state. Mitochondrial NAA production is dependent on functioning of respiratory chain enzyme complexes and oxidative phosphorylation. Inhibition of these enzymes leads to significant decrease in mitochondrial NAA and ATP production in isolated rat brain mitochondria (17).

54

This relates well with our observation of strong positive correlation between NAA content and respiratory chain complex II activity and NAA and ATP/ADP/AMP contents in PGLs. Further, we observed low NAA/total creatine ratio in tumours with compromised mitochondrial function (supplementary figure 3). Similar observations have also been made in patients with mitochondrial encephalomy-opathies (18). Thus, the NAA levels in SDH and VHL tumours are directly linked to the energetic states of the tumour cell mitochondria.

Figure 3 Purine metabolism pathway. Enzymes involved are: adenylosuccinate synthase (1), adenylosuccinate lyase (2), IMP dehydrogenase (3), adenosine deaminase (4), purine-nucleoside phosphorylase (5), xanthine dehydrogenase (6), hypoxanthine-guanine phosphoribosyl transferase (7), adenine phosphoribosyl transferase (8). PRPP, 5-phospho-a-D-ribosyl-1-pyrophosphate; SAICAR, succinylaminoimidazole carboxamide ribotide; AICAR, aminoimidazole carboxamide ribotide; FAICAR, formylaminoimidazole carboxamide ribotide; S-AMP, adenylosuccinate; S-adenosine, succinyladenosine. Figure adopted from Wevers et al., 1999 (15).

55

Genotype Specific Differences in PGL Metabolome

3

In our studies, we show the presence of high levels of lactate in some of the VHL tumours which also show low NAA and total creatine levels. In similar observations, 1H NMR spectroscopic imaging of rat gliomas in vivo also demonstrated reduced NAA levels in tumours associated with high glycolytic activity (19). Also, it has been observed that in tissues with ischaemic stroke there is increased anaerobic glycolysis as evidenced by increased levels of lactate in the lesion and the border zone brain tissue. Increased lactate levels in these tissues were associated with decreased NAA and total creatine levels. This was also associated with decreased glutamate and glutamine levels in the affected brain tissue (20). However, though we observed highly reduced NAA and reduced total creatine levels in tumours associated with increased aerobic glycolysis, the glutamate + glutamine levels were not different from other tumours. In SDH tumours, lower glutamate levels were observed compared to sporadic and VHL tumours by Imperiale et al. (21) using 1H high resolution magic angle spinning NMR spectroscopy in small tumour pieces. However, in the spectra of the present study obtained using homogenized tissue, glutamine and glutamate peaks were not clearly separated to verify the above mentioned report. Thus, estimation of levels of glutamate and glutamine with other methods might help in giving more clarity. Creatine serves as an energy reservoir and mode of energy transport within a cell by accepting a phosphate group in mitochondria through the activity of creatine kinase. Phosphorylated creatine can be shuttled to different cellular regions in need of high energy phosphate group to regenerate ATP. In cells with low energetic status, phosphocreatine levels are low. The reduction in total creatine levels as observed by us in VHL tumours is suggestive of larger alterations in cellular energy metabolism. In particular, it reflects a lack of sufficient ATP levels to maintain stores of phospho-creatine which might lead to efflux of cellular creatine stores. This might be reflected as low creatine levels in SDH and VHL tumours when compared to NF-1 tumours. Further, grouping of SDH tumours based on the tumour localization did not reveal location specific alterations in the metabolite levels. Thus, the tumour metabolism, as observed in the present study, tended to be influenced more in a genotype-specific way. In conclusion, alterations in energy metabolism caused by hereditary mutations are associated with alterations in catecholamine, amino acid and purine metabolism which could in addition to energy metabolism be targeted in these tumours for diagnostics and therapeutics.

56

3.5 References1. Lenders JW, Eisenhofer G, Mannelli M, Pacak K. Phaeochromocytoma. Lancet. 2005;366(9486):665-675.2. Welander J, Soderkvist P, Gimm O. Genetics and clinical characteristics of hereditary pheochromocytomas

and paragangliomas. Endocrine-related cancer. 2011;18(6):R253-276.3. Lorenzo FR, Yang C, Ng Tang Fui M, Vankayalapati H, Zhuang Z, Huynh T, Grossmann M, Pacak K,

Prchal JT. A novel EPAS1/HIF2A germline mutation in a congenital polycythemia with paraganglioma. Journal of molecular medicine. 2013;91(4):507-512.

4. Castro-Vega LJ, Buffet A, De Cubas AA, Cascon A, Menara M, Khalifa E, Amar L, Azriel S, Bourdeau I, Chabre O, Curras-Freixes M, Franco-Vidal V, Guillaud-Bataille M, Simian C, Morin A, Leton R, Gomez-Grana A, Pollard PJ, Rustin P, Robledo M, Favier J, Gimenez-Roqueplo AP. Germline mutations in FH confer predisposition to malignant pheochromocytomas and paragangliomas. Human molecular genetics. 2014.

5. Favier J, Briere JJ, Burnichon N, Riviere J, Vescovo L, Benit P, Giscos-Douriez I, De Reynies A, Bertherat J, Badoual C, Tissier F, Amar L, Libe R, Plouin PF, Jeunemaitre X, Rustin P, Gimenez-Roqueplo AP. The Warburg effect is genetically determined in inherited pheochromocytomas. PloS one. 2009;4(9): e7094.

6. Rao JU, Engelke UF, Rodenburg RJ, Wevers RA, Pacak K, Eisenhofer G, Qin N, Kusters B, Goudswaard AG, Lenders JW, Hermus AR, Mensenkamp AR, Kunst HP, Sweep FC, Timmers HJ. Genotype-specific abnormalities in mitochondrial function associate with distinct profiles of energy metabolism and catecholamine content in pheochromocytoma and paraganglioma. Clinical cancer research: an official journal of the American Association for Cancer Research. 2013;19(14):3787-3795.

7. Rapizzi E, Ercolino T, Canu L, Giache V, Francalanci M, Pratesi C, Valeri A, Mannelli M. Mitochondrial function and content in pheochromocytoma/paraganglioma of succinate dehydrogenase mutation carriers. Endocrine-related cancer. 2012;19(3):261-269.

8. Weljie AM, Jirik FR. Hypoxia-induced metabolic shifts in cancer cells: moving beyond the Warburg effect. The international journal of biochemistry & cell biology. 2011;43(7):981-989.

9. Wishart DS. Advances in metabolite identification. Bioanalysis. 2011;3(15):1769-1782.10. Weljie AM, Bondareva A, Zang P, Jirik FR. (1)H NMR metabolomics identification of markers of hypoxia -

induced metabolic shifts in a breast cancer model system. Journal of biomolecular NMR. 2011;49(3-4): 185-193.

11. Engelke UF, Kremer B, Kluijtmans LA, van der Graaf M, Morava E, Loupatty FJ, Wanders RJ, Moskau D, Loss S, van den Bergh E, Wevers RA. NMR spectroscopic studies on the late onset form of 3-methylgluta-conic aciduria type I and other defects in leucine metabolism. NMR in biomedicine. 2006;19(2):271-278.

12. Wevers RA, Engelke U, Wendel U, de Jong JG, Gabreels FJ, Heerschap A. Standardized method for high-resolution 1H-NMR of cerebrospinal fluid. Clin Chem. 1995;41(5):744-751.

13. Pillwein K, Chiba P, Knoflach A, Czermak B, Schuchter K, Gersdorf E, Ausserer B, Murr C, Goebl R, Stockhammer G, et al. Purine metabolism of human glioblastoma in vivo. Cancer Res. 1990;50(5):1576-1579.

14. Gottle M, Burhenne H, Sutcliffe D, Jinnah HA. Purine metabolism during neuronal differentiation: the relevance of purine synthesis and recycling. Journal of neurochemistry. 2013;127(6):805-818.

15. Wevers RA, Engelke UF, Moolenaar SH, Brautigam C, de Jong JG, Duran R, de Abreu RA, van Gennip AH. 1H-NMR spectroscopy of body fluids: inborn errors of purine and pyrimidine metabolism. Clinical chemistry. 1999;45(4):539-548.

16. Patel TB, Clark JB. Synthesis of N-acetyl-L-aspartate by rat brain mitochondria and its involvement in mitochondrial/cytosolic carbon transport. Biochem J. 1979;184(3):539-546.

17. Bates TE, Strangward M, Keelan J, Davey GP, Munro PM, Clark JB. Inhibition of N-acetylaspartate production: implications for 1H MRS studies in vivo. Neuroreport. 1996;7(8):1397-1400.

18. Mathews PM, Andermann F, Silver K, Karpati G, Arnold DL. Proton MR spectroscopic characterization of differences in regional brain metabolic abnormalities in mitochondrial encephalomyopathies. Neurology. 1993;43(12):2484-2490.

19. Ziegler A, von Kienlin M, Decorps M, Remy C. High glycolytic activity in rat glioma demonstrated in vivo by correlation peak 1H magnetic resonance imaging. Cancer Res. 2001;61(14):5595-5600.

57

Genotype Specific Differences in PGL Metabolome

3

20. van der Zijden JP, van Eijsden P, de Graaf RA, Dijkhuizen RM. 1H/13C MR spectroscopic imaging of regionally specific metabolic alterations after experimental stroke. Brain. 2008;131(Pt 8):2209-2219.

21. Imperiale A, Moussallieh FM, Sebag F, Brunaud L, Barlier A, Elbayed K, Bachellier P, Goichot B, Pacak K, Namer IJ, Taieb D. A new specific succinate-glutamate metabolomic hallmark in SDHx-related paragangliomas. PloS one. 2013;8(11):e80539.

58

3.6 Supplementary Methods

HPLC tandem MSThe HPLC tandem MS system consisted of a Quattro premier XE tandem MS (Micromass MS technologies, Waters, The Netherlands) in combination with an Acquity Ultraperformance LC (Waters, the Netherlands). Using the autosampler 5µl aliquots of prepared samples were injected on to an Atlantis T3 column (10cm, particle size 3µm). The mobile phase consisted of A: 0.3% formic acid in water and B: 100% methanol. Samples were eluted by running a step-wise gradient (99.5% buffer A from the start, 87.5% at 4.75min, 75% at 5.25min and 50% at 5.75min). Total run time amounted to 12.3min. The HPLC flow was directed to the tandem mass spectrometer fitted with an electrospray ionization probe operated in positive mode at unit resolution. The capillary voltage was set at 3kV. Settings for mass transitions and cone- and collision cell voltages were optimized for each component. Quantification was performed using an area based linear regression curve derived from the stable isotope diluted calibration standards.

Multivariate Statistical AnalysisTo build the data matrix, variable sized bucketing was performed between 10.00 – 0.50 ppm. The spectral regions between 5.2-4.5ppm, 3.8-3.3ppm and 1.19-1.15ppm were excluded from the bucket analysis to avoid the variation effects of water presaturation, methanol and ethanol residues, respectively. Signals showing larger chemical shift variations were compiled into broader buckets; for example alanine (bucket size 0.05; region 1.54 – 1.49 ppm). Bucket integrals were divided by the total spectral intensity. Bucketing was carried out in Amix version 3.9.14 (BRUKER BIOSPIN, Germany). To preserve natural differences in intensities and to highlight dominant effects, no scaling was performed before PCA. Principal Component Analysis (PCA) was performed using Unscrambler (CAMO Software AS, Oslo, Norway). PCA was performed without scaling to preserve natural differences in intensities and to highlight dominant effects. Further, to decrease the importance of the high intensities and find more subtle differences PCA was also performed after autoscaling (Supplementary figure 5).

Confirmation of metabolites identified by 1H-NMR spectroscopyConfirmation of the identified metabolite peak and validation of the concentrations as determined by 1H NMR spectroscopy were carried out for 1H NMR spectroscopic measurements of succinic acid and N-acetylaspartic acid (NAA) as their concentrations were measured by estimating the area under the curve for single peaks or singlets.

59

Genotype Specific Differences in PGL Metabolome

3

Correlation between metabolite concentrations as measured by 1H NMR spectroscopy and validated by other techniques was examined using Spearman’s correlation test and linear relationship was tested using Passing-Bablok regression analysis.

Determination of succinate by UPLC-MS/MS:Metabolite extraction: Fresh frozen tumour tissue (5-10 mg) was immersed in 500µl LC/MS grade methanol containing the internal standard 13C4-succinate (Catalogue No. 719269, Sigma-Aldrich, Germany). Homogenisation was achieved by vortexing with a metal bead for 2 minutes. Homogenates were transferred to a new tube and centrifuged at 2000xg for 5 minutes at 4°C to separate insoluble debris. Supernatants were taken to dryness with a speed vac concentrator (Thermo Scientific) and stored at -80°C. Pellets were resuspended in mobile phase and filtered using a centrifugal tube with 0.2 µm pore size. UPLC-MS/MS: UPLC-MS/MS analysis was carried out using an API QTRAP 5500 (AB Sciex) coupled to an Acquity UPLC system (Waters). Chromatographic separation was achieved on a Waters Acquity UPLC® HSS T3 column (1.8 µm, 2.1 x 100 mm) with a flow rate of 0.459 ml/min, a mobile phase of 0.2% formic acid in water (A) and 0.2% formic acid in acetonitrile (B), and a gradient of 0.37min 95% A, 4.87min 70% A, 5.37min 100% B, 5.87min 100% B, 6.37min 95% A, 7.5min 95% A. The column was maintained at a temperature of 25°C and the sample manager of 5°C. The mass spectrometer was operated at 550°C in negative ion electrospray mode. Multiple reaction monitoring was used for detection. Quantification and qualification of succinate was performed with the transitions 117/73 and 117/99, respectively. For the internal standard a transition of 121/76 was used. The Succinate content of the tumour tissue was validated using ultra performance liquid chromatography tandem MS (UPLC-MS/MS) for a subset of samples. A good and linear correlation was observed between two methods (supplementary fig 2).

Determination of N-acetyl aspartic acid by HPLC tandem MS:Deproteinized tumour tissue lysate recovered after 1H-NMR spectroscopy was used for measurement of N-acetyl aspartic acid. The samples were prepared by adding 50µl of stable isotope internal standard solution containing 10µM N-acetyl aspartic acid-d3 and 10µl of 2% formic acid for acidification to 50µl of deproteinized tumour tissue lysate. 5µl of the sample thus prepared was injected into the HPLC tandem MS system. Stock standards of 1mM of N-acetyl aspartic acid and N-acetyl aspartic acid -d3 were prepared in deionized water. The calibration standards of respectively 1, 2, 5, 10 and 20 µM of N-acetyl aspartic acid were prepared by diluting the stock solution with deionized water.

60

The HPLC-ms/ms system consisted of a Micromass Quattro LC connected to an Agilent HP1100 HPLC and a Gilson 232 XL autosampler. Using the autosampler 5µl aliquots of prepared sample were injected onto an Atlantis dC18 column (2.1mm X 100mm, particle size 3µm). The mobile phase consisted of A: 0.3% formic acid in water and B: 0.3% formic acid in 50% methanol/water. NAA and NAA-d3 were (co-)eluted by running a (linear) step gradient; starting at 1% B at a flow rate of 250 µl/min for 0.1 min, change to 100% B in 1.5 min. Column re-equilibration was initiated by increasing the flow rate to 280 µl/min in 0.1min, %B was kept at 100% for an additional minute and changed to 1% B in 0.1 min and kept at 1% B for 4.5 min. The flow rate was then decreased to the initial value of 250 µl/min in 0.3 min. The total run-time was 7.3 min. The autosampler wash solvent was water. The HPLC flow was directed to a Quattro LC (Micromass) tandem mass spectrometer fitted with an electrospray ionization probe operated in positive mode at unit resolution. The capillary voltage was set at 1 kV and a cone voltage of 14 V was used. The temperature settings for the source and ion block were respectively 400 ºC and 100 ºC. As a drying and nebulizing gas nitrogen was used at flow rates of respectively 650L/h and 100L/h. The collision cell was operated at 10 eV with Argon as collision gas at a pressure of 0.18 Pa. The mass spectrometer was set to monitor the ketene loss of both NAA and NAA-d3; recording the MRM mass transitions of m/z 176 to m/z 134 and m/z 179 to 137 m/z, respectively. Quantification was performed using an area based linear regression curve derived from the stable isotope diluted calibration standards.

61

Genotype Specific Differences in PGL Metabolome

3Supplementary Figure 1 Comparison of succinate and NAA content in tumour tissues as determined by HPLC and 1H-NMR spectroscopy. Passing - Bablok regression plot to compare measurement of epinephrine and norepinephrine content in PGL tumour tissues by HPLC and 1H-NMR spectroscopy expressed as µg per mg tissue in Log10 scale.

Supplementary Figure 2 PCA analysis of purines in PGLs tumours. A: The PCA scores plot shows the first and second principle components and clearly discriminate between SDH, RET and NF1 tumours. B: The metabolites contributing to the maximum variances in principal components 1 and 2 are shown outside the circles and in the outer circle of the loading plot indicating that the concentrations of adenine, hypoxanthine and inosine cause the separation between RET and SDH tumours in the PCA score plot.

A B

62

Supplementary Figure 3 Dot plot representing genotype specific differences in NAA/Creatine ratio. Horizontal line represents the mean. Data sets having different alphabets above them are significantly different with data sets having the alphabet ‘b’ being the lowest (p<0.05).

Supplementary Figure 4 Dot plots representing genotype specific differences in NAA and Creatine. Horizontal line represents the mean. Data sets having different alphabets above them are significantly different with data sets having the alphabet ‘a-d’ in the ascending order for value, ‘a’ being lowest to ‘d’ being highest (p<0.05).

N-acetyl aspartic Acid Creatine

b

a a

a, b, c

c

b, c

b,d

a,b,c,d

a a, c

a, b, c

d

63

Genotype Specific Differences in PGL Metabolome

3Supplementary Figure 5 PCA analysis of 1H NMR spectra of PPGLs with autoscaling. A: The PCA scores plot shows that the first and second principle components and clearly discriminate between SDH and RET tumours. B: The peak positions contributing to the maximum variances in principal components 1 and 2 are shown between the outer and inner circles of the loading plot.

A B

Overexpression of the natural antisense hypoxia-inducible factor-1α transcript is associated with malignant pheochromocytoma/paraganglioma

Span PN, Rao JU*, Oude Ophuis SBJ*, Lenders JWM, Sweep FCGJ, Wesseling P, Kusters B, van Nederveen FH, de Krijger RR, Hermus ARMM, Timmers HJLM

Endocr Relat Cancer. 2011 18(3):323-31.

* Authors contributed equally

4

66

Abstract

Paragangliomas (PGLs) have widely different metastatic potentials. Two different types of PGLs can be defined by expression profiling. Cluster 1 PGLs exhibit VHL and/or succinate dehydrogenase (SDH) mutations, and a pseudohypoxic phenotype. RET and neurofibromatosis type 1 (NF1) mutations occur in cluster 2 tumors characterized by deregulation of the RAS/RAF/MAP kinase signaling cascade. Sporadic PGLs can exhibit either profile. During sustained hypoxia, a natural antisense transcript of hypoxia inducible factor 1 (aHIF) is expressed. The role of aHIF in the metastatic potential of PGL has not yet been investigated.The aim was to test the hypothesis that genotype-specific overexpression of aHIF is associated with an increased metastatic potential. Tumor samples were collected from 87 patients with PGL. Quantitative PCR was performed for aHIF, VEGF, aquaporin 3, cytochrome b561, p57Kip2, slit homolog 3 and SDHC. Expression was related to mutation status, benign versus malignant tumors and metastasis free survival.We found that both aHIF and VEGF were overexpressed in cluster 1 PGLs, and in metastatic tumors. In contrast, slit homolog 3, p57Kip2, cytochrome b561 and SDHC showed overexpression in non-metastatic tumors, whereas no such difference was observed for aquaporin 3. Patients with higher expression levels of aHIF and VEGF had a significantly decreased metastasis-free survival. Higher expression levels of SDHC are correlated with an increased metastasis-free survival. In conclusion, we not only demonstrate a higher expression of VEGF in cluster 1 PGL, fitting a profile of pseudo-hypoxia and angiogenesis, but also of aHIF. Moreover, overexpression of aHIF and VEGF marks a higher metastatic potential in PGL.

67

aHIF in pheochromocytoma

4

4.1 Introduction

Paragangliomas (PGLs) are catecholamine-producing tumors that originate from chromaffin cells of the adrenal medulla (pheochromocytoma proper) or from sympathetic neuronal tissue in extra-adrenal locations of the abdomen or chest (1). PGLs in the head and neck usually derive from parasympathetic tissues. Metastases occur in 10-15% of patients with PGL (2). 25-30% of the PGLs are caused by germline mutations of one of the susceptibility genes (3), i.e. the proto-oncogene RET and the tumor-suppressor genes von Hippel-Lindau (VHL), neurofibromatosis type 1 (NF1) or succinate dehydrogenase subunit A, B, C and D (SDHA/B/C/D) (4). Mutations in the VHL and/or SDH genes result in a phenotype known as pseudo-hypoxia (5,6), leading to uncontrolled expression of HIF-1α regulated genes, such as vascular endothelial growth factor (VEGF) which in turn facilitates angiogenesis (7)(7). The signature of hypoxia and angiogenesis is not limited to VHL and/or SDH-related tumors, but is also prominent in a subset of sporadic PGLs, as shown by tumor gene expression profiling (8). These ‘cluster 1’ tumors are distinct from ‘cluster 2’ tumors, which include RET and NF1-related and another subset of sporadic PGLs (8). This latter cluster is characterized by the deregulation of the RAS/RAF/MAP kinase signaling cascade. High expression of HIF-1α and VEGF is associated with higher tumor aggressiveness and an unfavorable prognosis in many types of solid cancers (9,10). Hypoxia initially induces increased expression of HIF-1α, but during sustained hypoxia, the amount of HIF-1α mRNA is reduced (11). Instead, there is considerable overexpression of the HIF-1α natural antisense transcript (aHIF), which encodes the antisense template of the 3’ untranslated region of HIF-1α mRNA (12). This represents a negative feedback loop, in which aHIF inhibits the translation of HIF-1α mRNA after chronic hypoxia (13). A 20% of the human genome consists of these naturally occurring antisense transcripts, which regulate gene transcription, RNA splicing, polyadenylation, editing, stability, transport and translation (14-16). Importantly, besides acting as a transcription factor for hypoxia inducible genes, HIF-1α also stabilizes the tumor suppressor gene p53 under early hypoxic conditions (17). This situation is reversed during later stages of hypoxia, when HIF-1α translation is inhibited by aHIF (13), resulting in the loss of p53 followed by deregulated cell proliferation. Thus, aHIF expression also allows tumor cells to survive hypoxia without inducing the tumor suppressor p53 (12). Theoretically, aHIF expression is not mainly a phenotypical sign of chronic hypoxia, but also in itself mechanistically associated with more aggressive tumor behavior and increased metastatic potential (18). So far, the role of aHIF in the pathophysiology of PGL had not been investigated. We hypothesize that there is a genotype-specific overexpression of aHIF expression

68

in PGL, which is associated with an increased metastatic potential. To test this hypothesis, we performed quantitative RT-PCR of aHIF and other cluster-specific genes (VEGF, aquaporin 3, cytochrome b561, p57Kip2, slit homolog 3 and SDHC) in PGLs of different genotypes and investigated the relation between expression of these genes and metastasis-free survival (8).

4.2 Materials and methods

PatientsTotally, 151 patients with histologically proven pheochromocytoma or extra- adrenal sympathetic PGL evaluated at the Departments of Endocrinology and General Internal Medicine at the Radboud University Nijmegen Medical Centre (RUNMC) underwent surgical resection of primary PGL between November 1967 and November 2008. Details on the post-surgical follow-up of part of this cohort were published elsewhere (Timmers et al. 2008). Frozen primary tumor tissue was available from 87 patients (41 males, 46 females), 5 of whom underwent surgery at the Erasmus MC in Rotterdam. Tumors were adrenal in 76 cases and extra-adrenal abdominal in 11 cases. The presence of germline mutations and large deletions of RET, VHL, NF1 and SDHB/C/D was investigated at the Department of Genome Diagnostics of the University Medical Centre in Utrecht according to standard procedures. Clinical characteristics and genotypes are listed in Table 1. In 44 patients with an apparently sporadic presentation, i.e. no family history and lack of syndromal features, full germline testing of all susceptibility genes was omitted. In these cases, SDHB immunostaining was carried out on tumor tissues to investigate the functional integrity of SDH (see below). Mutations in SDHB, SDHC or SDHD genes lead to lack of SDHB immunoreactivity and thus can be used to reveal germline mutations in these genes (van Nederveen et al., 2009). During a mean±SD follow-up of 7.04 ± 5.07 (range: 0.05–20.83) years, 7 (8%) developed metastatic disease (Table 2). The mean ± SD interval between surgery and the diagnosis of metastases was 5.50 ± 8.15 (range: 0.00 – 19.38) years. Sixteen patients died, including 5 with metastatic PGL. Data were collected under conditions of regular clinical care, with the approval of ethics committee obtained for the use of those data for scientific purposes.

SDHB ImmunohistochemistryOf the 44 patients with apparently sporadic presentation, immunohistochemistry was performed in 37 patients to detect the expression of SDHB. Sections (4 µm) from paraffin-embedded tumor tissues were stained as described previously (van

69

aHIF in pheochromocytoma

4

Nederveen et al., 2009) using an anti-human rabbit polyclonal primary antibody (Sigma-Aldrich, St. Louis, MO, USA, HPA002868, 1:200).

Table 1 Clinical characteristics of paraganglioma (PGL) patients

Variable PGL patients (n=87)Age (years)Diagnosis mean ± SD 46.05 ± 16.17Diagnosis range 7 – 77Gender

Male 41Female 46Primary tumor locationAdrenal 76

Extra-adrenal 11Tumor size (cm)Mean ± SD 6.40 ± 3.42Range 1.4 – 18

GenotypeSDHB 3SDHD 1VHL 3MEN2A 10MEN2B 1NF1 7Sporadic* 18 (16 adrenal, 2 extra-adrenal)

Apparently sporadic 44 (39 adrenal, 5 extra-adrenal; 37/37 SDH-negativeb)Total follow-up (years)Mean ± SD 7.04 ± 5.07Range 0.05 – 20.83

MetastasesNo 80Yes 7

a. SDHB, SDHC, SDHD, and MEN2 mutations and large deletions ruled out germline testing.b. As assessed by positive SDHB immunohistochemistry

70

Tumor tissue processing Representative snap frozen tumor samples were cut in approximately five sections of 20 µm in a cryostat at -20°C, put in lysis buffer (RLT, Qiagen, Hilden, Germany) and stored at -80 °C until RNA isolation. For histological confirmation, additional slices were H&E stained for histological re-evaluation by a pathologist.

RNA extractionTotal RNA was extracted using the RNeasy Mini Kit (Qiagen) with on-column DNase-I treatment. RNA concentrations were determined by measuring the spectrophotometric absorption at 260 nm using the GeneQuant (Amersham, Eindhoven, The Netherlands). RT reactioncDNA synthesis was performed using 10 µl sample volume, containing approximately 1 µg total RNA and 10 µl RT-mix, with a total volume of 20 µl. cDNA synthesis was performed using a PTC-200 PCR apparatus (MJ Research). Samples were denatured for 10 minutes at 70°C and immediately cooled on ice. After random hexamers were annealed for 10 minutes at 21 °C, cDNA synthesis was performed for 45 minutes at 42 °C, followed by an enzyme inactivation step for 5 minutes at 99 °C.

Quantitative PCRFor the quantitative PCR (qPCR), SYBR Green assays were used for aHIF and VEGF, and Taqman for aquaporin 3 (Gill blood group), cytochrome b561, p57Kip2, slit homolog 3 (Drosophila melanogaster) and SDHC. The latter 5 genes were selected because in a previous study by Dahia et al. (8), their RNA expression was shown to mark either cluster 1 or 2 PGL. Overexpression of VEGF and aquaporin 3 mark cluster 1 tumors, whereas Cytochrome b561, P57Kip2, Slit homolog 3 and SDHC mark cluster 2 tumors. qPCR was performed using the Taqman Universal PCR Master Mix or SYBR Green Universal Master Mix (PE Applied Biosystems, Nieuwerkerk aan den IJssel, The Netherlands) in a total volume of 25 µl. The forward and reverse primers, as well as the optimal concentrations (for each target gene), were optimized in previous studies at the Department of Laboratory Medicine of the RUNMC for the SYBR Green assays, whereas Assays on Demand (PE Applied Biosystems) was used for the Taqman based assays. Hypoxan-thine-guanine phosphoribosyltransferase (HPRT1) was used as a housekeeping gene for normalization of all samples (19). qPCR was performed using the ABI Prism 7700 Sequence Detector (PE Applied Biosystems). The amplification reactions were carried out with denaturation at 95°C for 10 min, 40 cycles of 15 s at 95°C (melting) and 60 s at 60°C (annealing and elongation).

71

aHIF in pheochromocytoma

4

Statistical analysisStatistical analysis was performed using Statistical Package for the Social Sciences (for Windows 16.0; SPSS 16.0, Chicago, IL, USA). Aquaporin-3 mRNA levels were normally distributed (one-sample Kolmogorov-Smirnov), whereas those of other genes were not. After log-transformation a normal distribution of mRNA expression was found for the other genes. Thus, parametric tests were used for all the log- transformed mRNA levels, with the exception of Aquaporin-3, which was already normally distributed before log-transformation. Analysis of Variance (ANOVA) with post-hoc Tukey HSD test was used to test for the differences in mRNA expression of the target genes between cluster 1 and 2 PGLs, sporadic PGLs and apparently sporadic PGLs with unknown genotype. Differences in the mRNA expression between metastatic and non-metastatic tumors were tested using an independent samples t-test. In addition, the relation between expression of individual genes and meta-stasis-free survival was investigated. Normalized mRNA levels of an individual target gene were correlated with metastasis-free survival after dichotomizing on the basis of the median mRNA expression of individual genes using univariate Cox regression survival analysis. Metastasis-free survival of patients with an above versus below median expression was compared using the Kaplan Meier method, and a log-rank test for equality of survival distributions. A similar analysis was performed for overall survival.

4.3 Results

SDHB expression in apparently sporadic tumor tissuesPresence of SDHB immunoreactivity and a granular staining pattern typical of mitochondrial proteins was detected in all the tumor tissues of apparently sporadic cases. This includes the four metastatic cases with incomplete germline mutation testing (supplementary figure).

Genotype-specific mRNA expressionFigure 1 illustrates differences in mRNA expression of the individual target genes between sporadic PGLs (n = 18), apparently sporadic PGLs (n = 44), and cluster 1 PGLs (SDHB (n = 3), SDHD (n = 1), VHL (n = 3)) and cluster 2 PGLs (MEN2A (n = 10), MEN2B (n = 1), NF1 (n = 7)). After performing an ANOVA with post-hoc Tukey HSD test to compare mRNA expression of individual genes, both VEGF (figure 1A) and aHIF (figure 1B) expression was found to be significantly higher in cluster 1 as compared to cluster 2 PGLs (P = 0.018 and P = 0.021, respectively). Sporadic and apparently sporadic tumors exhibited expression levels that were not significantly different from cluster 1 and cluster 2 tumors. No differences were found in

72

aquaporin 3, cytochrome b561, p57Kip2, slit homolog 3 or SDHC expression (not shown).

To further explore the significantly higher expression of VEGF and aHIF in cluster 1 compared to cluster 2 PGLs, genotype-specific mRNA expression levels are shown in figure 2. SDHB, SDHD and VHL (cluster 1) tumors on average exhibit higher expression levels for both aHIF and VEGF than MEN2A, MEN2B and NF1 (cluster 2) related tumors. The size of the subgroups precluded relevant statistical analyses.

mRNA expression and metastasis-free survival The mRNA expression of each gene in non-metastatic PGLs (n = 80) versus the mRNA expression in the primary tumors of metastatic PGLs (n = 7) is visualized in figure 3. The RNA expression of both VEGF (P = 0.001) and aHIF (P = 0.001) was significantly higher in the primary tumors of PGLs that metastasized during follow up, than in non-metastatic PGLs (figure 3a-b). Aquaporin 3 expression was not significantly different between groups (figure 3c), whereas the expression of slit homolog 3 (P = 0.019), p57KIP2 (P = 0.035), cytochrome b561 (P = 0.003) and SDHC (P = 0.039) was significantly higher in the non-metastatic group than in the metastatic group (figure 3d-g). Subsequently, we assessed whether metastasis-free survival differed between the groups of patients after dichotomization at the median expression level using a Kaplan-Meier analysis with log-rank testing. This analysis indicates that high

Figure 1 Box and whiskers plots (box= median with 25th and 75th percentile, and whiskers are minimum and maximum values) of log normalized expression levels of VEGF (A) and aHIF (B) in sporadic, apparently sporadic, and cluster 1 or cluster 2 type of PGLs. Significant differences, as assessed by ANOVA and post-hoc Tukey HSD testing, were found only between cluster 1 and cluster 2 tumors.

A B

73

aHIF in pheochromocytoma

4expression of VEGF is associated with decreased metastasis-free survival (figure 4a, P = 0.004). This is also true for high aHIF expression (Figure 4b, P = 0.011). Among the seven metastatic cases, high VEGF and aHIF expression did not seem to correlate with (metastasis-free) survival (data not shown). An opposite effect was observed for SDHC expression (P = 0.008) (figure 4g). For the other genes, metasta-sis-free survival was not significantly different between the high versus low expression groups, although low levels of expression were without exception associated with poor prognosis. Overall survival was not significantly different between the high versus low expression groups, although there was trend towards a higher mortality in patients with high VEGF expression (P = 0.067, data not shown).

Discussion

This is the first study in which the role of the natural antisense transcript aHIF is investigated in PGL. We found that aHIF, along with VEGF, is overexpressed in PGLs with a pseudohypoxic signature due to SDH and VHL mutations. Over- expression of both factors is also associated with a decreased metastasis-free survival, marking an increased malignant potential of apparently benign primary PGLs. Conversely, slit homolog 3, p57Kip2, cytochrome b561 and SDHC showed overexpression in non-metastatic tumors, whereas no difference was observed for aquaporin 3. Higher expression levels of SDHC are correlated with an increased metastasis-free survival.

Figure 2 Box and whiskers plots (box= median with 25th and 75th percentile, and whiskers are minimum and maximum values) of log normalized expression levels of VEGF (A) and aHIF (B) in sporadic, apparently sporadic, and PGLs defined by specific genotype.

A B

74

HIF-1α regulates the transcription of a number of genes that are known to be involved in tumorigenesis and angiogenesis (20), including VEGF (21). In VHL disease, stabilization of HIF-1α in PGL tumor cells is the result of decreased pVHL-dependent proteosomal degradation of HIF-1α by the E3 ubiquitin ligase

Figure 3 Box and whiskers plots (box= median with 25th and 75th percentile, and whiskers are minimum and maximum values) of expression levels of VEGF (A), aHIF (B), Aquaporin-3 (C), SLIT3 (D), KIP2 (E), Cytochrome b561 (F) and SDHC (G) between non-metastatic and metastatic PGL tumors. Significances are for Student’s t-testing of normalized values.

A

C

E

G

B

D

F

75

aHIF in pheochromocytoma

4

complex, after being hydroxylated by prolyl hydroxylase (6). In case of underlying mutations of the mitochondrial enzyme SDH, succinate accumulation inhibits prolyl hydroxylase activity and consequently HIFα stabilization (20). As shown in

Figure 4 Kaplan-Meier plots of metastasis free survival in PGL patients dichotomized by the mean tumor expression level of VEGF (A), aHIF (B), Aquaporin-3 (C), SLIT3 (D), KIP2 (E), Cytochrome b561 (F) or SDHC (G). Straight lines > median, dashed lines < median. Significances are for log-rank testing.

A

C

E

G

B

D

F

76

SDHA, B and D-associated PGL, abolishment of SDH activity results in over- expression of VEGF, a target gene of HIF-1α (5,22,23). Mitochondrial dysfunction due to SDH-mutations and activation of the hypoxic pathway is probably interlinked with other mechanisms of tumorigenesis involving oxidative stress and apoptosis resistance (24). Reactive oxygen species were shown to inhibit the action of prolyl hydroxylase, thereby increasing HIF-1α and VEGF availability (25). Another important theory interlinking the pathogenesis of different hereditary forms of PGL pertains to the developmental culling of sympathetic neuronal precursor cells by affecting c-Jun-dependent neuronal apoptosis following withdrawal of neuronal growth factor (26).The role of hypoxia and angiogenesis as a pathophysiological mechanism in a subset of PGLs was confirmed by gene expression profiling studies by Dahia and coworkers (8). PGLs caused by germline VHL and SDH mutations exhibit a clear signature of (pseudo) hypoxia and angiogenesis, assigned as ‘cluster 1’ tumors. On the other hand, RET and NF1 related PGLs, so-called cluster 2 tumors, are characterized by a deregulated RAS/RAF/MAP kinase signaling cascade which is associated with metabolism, translation initiation and RNA/protein synthesis. Subsets of sporadic tumors were shown to fit either the cluster 1 or 2 profile. Our current findings of overexpression of VEGF in cluster 1 tumors and corroborate these previous findings. The expression of other genes that were suggested to be cluster-related, i.e. p57Kip2, slit homolog 3, aquaporin 3, Cytochrome b561 and SDHC, was not significantly different between the groups in our study. Besides by factors mentioned above, the expression of HIF-1α is also regulated by aHIF, a naturally occurring antisense transcript RNA that forms a double stranded RNA molecule with the antisense transcript of HIF-1α (12,13). Naturally occurring antisense transcript expression can lead to the down-regulation of the sense transcript by RNA interference (14,16). Naturally occurring antisense transcript RNAs also exhibit important biological functions in DNA methylation (27), genomic imprinting (28), X-chromosome inactivation (29), genomic recombination (30), and RNA splicing, polyadenylation, editing, stability, transport and translation (31) (32). In vitro experiments in renal cell carcinoma cells have yielded insight in the oxygen-dependent regulation of HIF-1α (12,13). Hypoxic conditions initially induce overexpression of HIF-1α. During sustained hypoxia, however, HIF-1α mRNA decreases in parallel with increased expression of aHIF, suggesting inhibition of HIF-1α by aHIF through a negative feedback loop. Our current finding of increased aHIF expression in parallel with VEGF in SDH- and VHL-related PGL fits a pattern of a chronic (pseudo-) hypoxic state. This suggests that aHIF can serve as a surrogate marker of chronic HIF-1α activation, taking into account that measurement of the HIF-1α protein itself is cumbersome.

77

aHIF in pheochromocytoma

4

Besides its involvement in hypoxia-related tumorigenesis, HIF-1α stabilizes the tumor suppressor gene p53 (17). Inhibition of HIF-1α by aHIF during sustained hypoxia results in the loss of p53 and subsequent tumor cell proliferation (13). This aHIF-mediated escape from p53-mediated apoptosis theoretically translates into more aggressive tumor behavior and increased metastatic potential. Over- expression of aHIF has been clearly established in malignant non-papillary clear cell renal carcinoma (12). In breast cancer, overexpression of aHIF was shown to be a marker of poor prognosis (18). In the present study, aHIF expression in primary tumors turned out to be a predictor of decreased metastasis-free survival. There are currently no reliable histological criteria for malignant PGL besides the presence of metastatic lesions in locations where chromaffin tissue is normally absent (33). The 5-year survival of patients with metastatic PGL is approximately 50% (34,35), and no cure is available. It is therefore critical to identify biomarkers of malignancy such as aHIF. Besides aHIF, VEGF was also negatively correlated with metastasis-free survival. High expression of Cytochrome b561, p57Kip2, slit homolog 3 and SDHC on the other hand was associated with a benign course. These expression profiles of malignant PGL fit those of cluster 1 tumors. Among cluster 1 tumors, specifically SDHB mutations are strongly associated with a malignant course, whereas VHL- and SDHD-related PGLs are rarely malignant. Our findings, however, probably do not specifically reflect the SDHB genotype, since only two out of seven patients who developed metastases were known mutation carriers. However, considering the low frequency of malignancy in VHL/SDHD-related tumors, a cluster 1 signature of hypoxia/angiogenesis along with increased aHIF and VEGF expression as observed in our VHL and SDHD samples does not automatically translate into a high malignant potential. The discrepancy in metastatic potential between cluster 1 tumors of different genotypes suggests that other pathophysiological mechanisms besides hypoxia/angiogenesis are involved. Also, we emphasize that the expression of aHIF is highly variable, so it cannot be used as a marker of malignancy in individual patients. It is likely that individual risks of developing metastases can only be estimated by use of a combination of clinical, biochemical, genetic and molecular markers (36,37). In this context, protein expression of SNAIL, a zinc-finger transcription factor, has been identified as a useful marker of metastatic potential of PGL (38). Our current study has several limitations. The study includes a limited number of patients with metastatic disease. The role of aHIF as a predictor of malignancy needs to be further evaluated in a larger cohort with a longer follow-up. Also, the role of aHIF as a marker of the up-regulation of hypoxic and angiogenic pathways in the tumorigenesis of PGL needs further exploration across tumors of different genotypes, specifically SDHB versus other cluster 1

78

genotypes. Since the samples that we investigated spanned several decades, the integrity of mRNA could theoretically have decreased over time in the older samples. We did, however, not find such a correlation for aHIF expression. Formal genetic testing was omitted in several patients with apparently sporadic PGL, i.e. in those with a negative family history and lack of syndromal features. In these cases, SDHB immunostaining was used as ‘surrogate’ genotyping. The sensitivity and specificity of this approach awaits prospective evaluation. In conclusion, along with VEGF, the expression of aHIF, a naturally occurring antisense transcript of HIF-1α, is increased in cluster 1 PGL, fitting a profile of pseudo-hypoxia and angiogenesis. Moreover, overexpression of aHIF and VEGF marks a higher metastatic potential in PGL.

79

aHIF in pheochromocytoma

4

4.5 References1. De Lellis RA, Lloyd RV, Heitz PU, Eng C. Pathology and Genetics: World Health Organization classification

of tumours of endocrine organs. . Oxford, UK: Oxford University Press; 2004.2. Lenders JW, Eisenhofer G, Mannelli M, Pacak K. Phaeochromocytoma. Lancet. 2005;366(9486):665-675.3. Gimenez-Roqueplo AP, Burnichon N, Amar L, Favier J, Jeunemaitre X, Plouin PF. Recent advances in

the genetics of phaeochromocytoma and functional paraganglioma. Clinical and experimental pharmacology & physiology. 2008;35(4):376-379.

4. Baysal BE. Clinical and molecular progress in hereditary paraganglioma. Journal of medical genetics. 2008;45(11):689-694.

5. Gimenez-Roqueplo AP, Favier J, Rustin P, Mourad JJ, Plouin PF, Corvol P, Rotig A, Jeunemaitre X. The R22X mutation of the SDHD gene in hereditary paraganglioma abolishes the enzymatic activity of complex II in the mitochondrial respiratory chain and activates the hypoxia pathway. Am J Hum Genet. 2001;69(6):1186-1197.

6. Kaelin WG. Von Hippel-Lindau disease. Annual review of pathology. 2007;2:145-173.7. Stoppa-Lyonnet D, Lenoir G. [Cancer genetic predisposition: current events and perspectives 2005].

Med Sci (Paris). 2005;21(11):962-968.8. Dahia PL, Ross KN, Wright ME, Hayashida CY, Santagata S, Barontini M, Kung AL, Sanso G, Powers JF,

Tischler AS, Hodin R, Heitritter S, Moore F, Dluhy R, Sosa JA, Ocal IT, Benn DE, Marsh DJ, Robinson BG, Schneider K, Garber J, Arum SM, Korbonits M, Grossman A, Pigny P, Toledo SP, Nose V, Li C, Stiles CD. A HIF1alpha regulatory loop links hypoxia and mitochondrial signals in pheochromocytomas. PLoS genetics. 2005;1(1):72-80.

9. Hoogsteen IJ, Marres HA, Bussink J, van der Kogel AJ, Kaanders JH. Tumor microenvironment in head and neck squamous cell carcinomas: predictive value and clinical relevance of hypoxic markers. A review. Head & neck. 2007;29(6):591-604.

10. Manders P, Sweep FC, Tjan-Heijnen VC, Geurts-Moespot A, van Tienoven DT, Foekens JA, Span PN, Bussink J, Beex LV. Vascular endothelial growth factor independently predicts the efficacy of postoperative radiotherapy in node-negative breast cancer patients. Clinical cancer research : an official journal of the American Association for Cancer Research. 2003;9(17):6363-6370.

11. Wenger RH, Rolfs A, Spielmann P, Zimmermann DR, Gassmann M. Mouse hypoxia-inducible factor-1alpha is encoded by two different mRNA isoforms: expression from a tissue-specific and a housekeeping-type promoter. Blood. 1998;91(9):3471-3480.

12. Thrash-Bingham CA, Tartof KD. aHIF: a natural antisense transcript overexpressed in human renal cancer and during hypoxia. Journal of the National Cancer Institute. 1999;91(2):143-151.

13. Uchida T, Rossignol F, Matthay MA, Mounier R, Couette S, Clottes E, Clerici C. Prolonged hypoxia differentially regulates hypoxia-inducible factor (HIF)-1alpha and HIF-2alpha expression in lung epithelial cells: implication of natural antisense HIF-1alpha. The Journal of biological chemistry. 2004;279(15):14871-14878.

14. Chen J, Sun M, Kent WJ, Huang X, Xie H, Wang W, Zhou G, Shi RZ, Rowley JD. Over 20% of human transcripts might form sense-antisense pairs. Nucleic acids research. 2004;32(16):4812-4820.

15. Katayama S, Tomaru Y, Kasukawa T, Waki K, Nakanishi M, Nakamura M, Nishida H, Yap CC, Suzuki M, Kawai J, Suzuki H, Carninci P, Hayashizaki Y, Wells C, Frith M, Ravasi T, Pang KC, Hallinan J, Mattick J, Hume DA, Lipovich L, Batalov S, Engstrom PG, Mizuno Y, Faghihi MA, Sandelin A, Chalk AM, Mottagui-Tabar S, Liang Z, Lenhard B, Wahlestedt C, Group RGER, Genome Science G, Consortium F. Antisense transcription in the mammalian transcriptome. Science. 2005;309(5740):1564-1566.

16. Yelin R, Dahary D, Sorek R, Levanon EY, Goldstein O, Shoshan A, Diber A, Biton S, Tamir Y, Khosravi R, Nemzer S, Pinner E, Walach S, Bernstein J, Savitsky K, Rotman G. Widespread occurrence of antisense transcription in the human genome. Nature biotechnology. 2003;21(4):379-386.

17. An WG, Kanekal M, Simon MC, Maltepe E, Blagosklonny MV, Neckers LM. Stabilization of wild-type p53 by hypoxia-inducible factor 1alpha. Nature. 1998;392(6674):405-408.

80

18. Cayre A, Rossignol F, Clottes E, Penault-Llorca F. aHIF but not HIF-1alpha transcript is a poor prognostic marker in human breast cancer. Breast cancer research : BCR. 2003;5(6):R223-230.

19. de Kok JB, Roelofs RW, Giesendorf BA, Pennings JL, Waas ET, Feuth T, Swinkels DW, Span PN. Normalization of gene expression measurements in tumor tissues: comparison of 13 endogenous control genes. Laboratory investigation; a journal of technical methods and pathology. 2005;85(1):154-159.

20. Selak MA, Armour SM, MacKenzie ED, Boulahbel H, Watson DG, Mansfield KD, Pan Y, Simon MC, Thompson CB, Gottlieb E. Succinate links TCA cycle dysfunction to oncogenesis by inhibiting HIF-alpha prolyl hydroxylase. Cancer Cell. 2005;7(1):77-85.

21. Forsythe JA, Jiang BH, Iyer NV, Agani F, Leung SW, Koos RD, Semenza GL. Activation of vascular endothelial growth factor gene transcription by hypoxia-inducible factor 1. Molecular and cellular biology. 1996;16(9):4604-4613.

22. Burnichon N, Briere JJ, Libe R, Vescovo L, Riviere J, Tissier F, Jouanno E, Jeunemaitre X, Benit P, Tzagoloff A, Rustin P, Bertherat J, Favier J, Gimenez-Roqueplo AP. SDHA is a tumor suppressor gene causing paraganglioma. Hum Mol Genet. 2010;19(15):3011-3020.

23. Gimenez-Roqueplo AP, Favier J, Rustin P, Rieubland C, Kerlan V, Plouin PF, Rotig A, Jeunemaitre X. Functional consequences of a SDHB gene mutation in an apparently sporadic pheochromocytoma. The Journal of clinical endocrinology and metabolism. 2002;87(10):4771-4774.

24. Gottlieb E, Tomlinson IP. Mitochondrial tumour suppressors: a genetic and biochemical update. Nature reviews Cancer. 2005;5(11):857-866.

25. Pouyssegur J, Mechta-Grigoriou F. Redox regulation of the hypoxia-inducible factor. Biological chemistry. 2006;387(10-11):1337-1346.

26. Lee S, Nakamura E, Yang H, Wei W, Linggi MS, Sajan MP, Farese RV, Freeman RS, Carter BD, Kaelin WG, Jr., Schlisio S. Neuronal apoptosis linked to EglN3 prolyl hydroxylase and familial pheochromocyto-ma genes: developmental culling and cancer. Cancer Cell. 2005;8(2):155-167.

27. Tufarelli C, Stanley JA, Garrick D, Sharpe JA, Ayyub H, Wood WG, Higgs DR. Transcription of antisense RNA leading to gene silencing and methylation as a novel cause of human genetic disease. Nature genetics. 2003;34(2):157-165.

28. Rougeulle C, Heard E. Antisense RNA in imprinting: spreading silence through Air. Trends Genet. 2002;18(9):434-437.

29. Ogawa Y, Lee JT. Antisense regulation in X inactivation and autosomal imprinting. Cytogenetic and genome research. 2002;99(1-4):59-65.

30. Bolland DJ, Wood AL, Johnston CM, Bunting SF, Morgan G, Chakalova L, Fraser PJ, Corcoran AE. Antisense intergenic transcription in V(D)J recombination. Nature immunology. 2004;5(6):630-637.

31. Mello CC, Conte D, Jr. Revealing the world of RNA interference. Nature. 2004;431(7006):338-342.32. Makalowska I, Lin CF, Makalowski W. Overlapping genes in vertebrate genomes. Computational

biology and chemistry. 2005;29(1):1-12.33. Linnoila RI, Keiser HR, Steinberg SM, Lack EE. Histopathology of benign versus malignant sympath-

oadrenal paragangliomas: clinicopathologic study of 120 cases including unusual histologic features. Human pathology. 1990;21(11):1168-1180.

34. Eisenhofer G, Bornstein SR, Brouwers FM, Cheung NK, Dahia PL, de Krijger RR, Giordano TJ, Greene LA, Goldstein DS, Lehnert H, Manger WM, Maris JM, Neumann HP, Pacak K, Shulkin BL, Smith DI, Tischler AS, Young WF, Jr. Malignant pheochromocytoma: current status and initiatives for future progress. Endocrine-related cancer. 2004;11(3):423-436.

35. Timmers HJ, Pacak K, Huynh TT, Abu-Asab M, Tsokos M, Merino MJ, Baysal BE, Adams KT, Eisenhofer G. Biochemically silent abdominal paragangliomas in patients with mutations in the succinate dehydrogenase subunit B gene. The Journal of clinical endocrinology and metabolism. 2008;93(12):4826-4832.

36. Eisenhofer G, Huynh TT, Pacak K, Brouwers FM, Walther MM, Linehan WM, Munson PJ, Mannelli M, Goldstein DS, Elkahloun AG. Distinct gene expression profiles in norepinephrine- and epineph-rine-producing hereditary and sporadic pheochromocytomas: activation of hypoxia-driven angiogenic pathways in von Hippel-Lindau syndrome. Endocrine-related cancer. 2004;11(4):897-911.

81

aHIF in pheochromocytoma

4

37. Suh I, Shibru D, Eisenhofer G, Pacak K, Duh QY, Clark OH, Kebebew E. Candidate genes associated with malignant pheochromocytomas by genome-wide expression profiling. Annals of surgery. 2009;250(6): 983-990.

38. Hayry V, Salmenkivi K, Arola J, Heikkila P, Haglund C, Sariola H. High frequency of SNAIL-expressing cells confirms and predicts metastatic potential of phaeochromocytoma. Endocrine-related cancer. 2009;16(4):1211-1218.

82

Supplementary Figure 1 SDHB immunostaining in pheochromocytoma tumor tissues from patients with metastatic disease. Panel A: 58 year old male (Table 2), right adrenal PGL. Panel B: 40 year old male (Table 2), left adrenal PGL. An image of higher magnification (20X) is enclosed in the inset. Note: Strong granular staining pattern in the tumor cells.

A B

83

aHIF in pheochromocytoma

4

Correlation between in vivo 18F-fluorodeoxyglucose PET and immunohistochemical markers of glucose uptake and metabolism in pheochromocytoma and paraganglioma

Rao JU*, van Berkel A*, Kusters B, Demir T, Visser E, Mensenkamp AR, van der Laak JAWM, Oosterwijk E, Lenders JWM, Sweep FCGJ, Wevers RA, Hermus ARMM, Langenhuijsen JF, Kunst HPM, Pacak K, Gotthardt M, Timmers HJLM.

J Nucl Med. 2014 55(6):1253-59.

* Authors contributed equally

5

86

Abstract

Pheochromocytomas and paragangliomas (PPGLs) can be localized by 18F-FDG PET. The uptake is particularly high in tumours with an underlying succinate dehydrogenase (SDH) mutation. The latter are characterized by compromised oxidative phosphorylation and a pseudo-hypoxic response which mediates an increase in aerobic glycolysis, also known as the Warburg effect. The aim of this study was to explore the hypothesis that increased uptake of 18F-FDG in SDHx-related PPGLs is reflective of increased glycolytic activity and is correlated with expression of different proteins involved in glucose uptake and metabolism through the glycolytic pathway.Methods: Twenty-seven PPGLs collected from patients with hereditary mutations in SDHB (n=2), SDHD (n=3), RET (n=5), NF1 (n=1), MAX (n=1) and sporadic patients (n=15) were investigated. Pre-operative 18F-FDG PET/CT studies were analyzed; mean and maximum standardized uptake values (SUVs) in manually drawn regions of interest were calculated. Expression of proteins involved in glucose uptake (GLUT-1 and 3), phosphorylation (HK-1, 2 and 3), glycolysis (MCT-4) and angiogenesis (VEGF, CD34) were examined in paraffin-embedded tumour tissues using immunohistochemical staining with peroxidase-catalyzed polymerization of diaminobenzidine as read-out. Expression was correlated with corresponding SUVs.Results: Both maximum and mean SUVs for SDHx-related tumours were significantly higher than for sporadic and other hereditary tumours (P<0.01). The expression of HK-2 and HK-3 were significantly higher in SDHx-related PPGLs compared to sporadic PPGLs (P=0.022, P=0.025). The expression of HK-2 and VEGF were significantly higher in SDHx-related PPGLs compared to other hereditary PPGLs (P=0.039, P=0.008). No statistical differences in the expression were observed for GLUT-1, GLUT-3 and MCT-4. Percentage anti-CD 34 staining and mean vessel perimeter were significantly higher in SDHx-related PPGLs compared to sporadic tumours (P=0.050, P=0.010). Mean SUVs significantly correlated with the expression of HK-2 (P=0.027), HK-3 (P=0.013), VEGF (P=0.049) and MCT-4 (P=0.020). Conclusion: Activation of aerobic glycolysis in SDHx-related PPGLs results in increased 18F-FDG accumulation and is associated with accelerated glucose phos-phorylation by hexokinases rather than increased expression of glucose transporters.

87

18F-FDG PET in PPGL

5

5.1 Introduction

Pheochromocytomas and paragangliomas (PPGLs) are catecholamine-producing neuroendocrine tumours arising from the chromaffin cells of the adrenal medulla and extra-adrenal sympathetic paraganglia. PPGLs can occur as part of hereditary syndromes. Susceptibility genes include succinate dehydrogenase (SDH) complex subunits and assembly factor 2 (SDHA/B/C/D/AF2), von Hippel-Lindau (VHL), RET, neurofibromatosis (NF) 1, myc-associated factor X (MAX) and transmembrane protein 127 (TMEM127) (1). Gene expression profiling has revealed the presence of two clusters of PPGLs: cluster 1 (VHL, SDHx) which show increased expression of genes associated with angiogenesis and hypoxia whereas cluster 2 (RET, NF1, TMEM127 and MAX) display a rich signature of RNA synthesis and kinase signaling (2,3). Enhanced uptake of glucose by tumour cells compared to normal cells is the hallmark of in vivo cancer imaging with 18F-fluorodeoxyglucose positron emission computed tomography (18F-FDG PET/CT). We have shown previously that 18F-FDG PET/CT is superior to other functional imaging techniques for localizing metastatic PPGL, particularly in those with an underlying SDHB mutation (4-6). 18F-FDG PET/CT is also very useful for localizing benign PPGLs (7). Interestingly, 18F-FDG uptake varies among PPGLs of different genotypes, with the highest standard uptake values (SUVs) being observed in SDH and VHL-related tumours (5,8). The precise mechanism behind these genotype-specific differences in 18F-FDG uptake has not been elucidated. SDHx mutations cause impairment of SDH function in the mitochondrial electron transport chain and hence compromise oxidative phosphorylation (9-11). Abolition of SDH enzymatic activity results in activation of the hypoxic- angiogenic pathway via transcription factor hypoxia-inducible factors (HIFs)-1α and -2α (12). Their main target genes include genes involved in glucose metabolism (glucose transporters (GLUT), hexokinases (HK), angiogenesis (vascular endothelial growth factor (VEGF)), survival and motility (13-15). Activation of HIF-α further supports the shift of tumour cell energy metabolism from oxidative phosphoryla-tion towards aerobic glycolysis, also known as the Warburg effect (16). The alternative energy generation pathway is somewhat less efficient requiring a much larger cellular influx of glucose to maintain tumour cells energy needs. High uptake of 18F-FDG by SDHx-related tumours has been suggested to be a reflection of the Warburg effect. Various mechanisms for accelerated glucose use by tumour cells have been described. Enhanced influx of glucose via glucose transporters (GLUTs) is considered to be the most important. Overexpression of GLUT isoforms GLUT-1 and 3 is closely related to 18F-FDG uptake in tumour cells (17). In addition, accelerated glucose phosphorylation by the cytosolic enzyme hexokinase (HK) as the first step toward glycolysis results in enhanced 18F-FDG

88

accumulation. HK-2 is predominantly expressed in tumour cells that exhibit the Warburg effect (18) and is associated with elevated 18F-FDG uptake in malignant conditions (19,20). Upregulation of both GLUTs and HK is frequently associated with malignant transformation of cells (21). Furthermore, activity of HK-3 and monocarboxylate transporter type 4 (MCT-4), which facilitates the cellular lactate transport, possibly are regulated by hypoxia (10,22). In addition, hypoxia also promotes anaerobic glycolysis and several studies have demonstrated that 18F-FDG uptake is an indirect reflection of tumour hypoxia (23,24). The aim of this study was to explore the hypothesis that increased uptake of 18F-FDG is reflective of increased glycolytic activity and is correlated with expression of different proteins involved in glucose uptake and metabolism through the glycolytic pathway. Therefore, immunohistochemical expression of GLUT-1, GLUT-3, HK-1, HK-2, HK-3 and MCT-4 was directly correlated with in vivo 18F-FDG uptake in PPGLs of different genotypes. Additionally, VEGF expression was examined to account for hypoxia regulated angiogenesis, and also CD34 to account for genotype-specific differences in microvasculature that could alter the radiotracer supply to tumour cells.

5.2 Materials and methods

PatientsThe study included 27 consecutive patients (17 men and 10 women, aged [mean±SD] 51±15 years) in whom PPGL was histopathologically confirmed and of which 22 (81%) were adrenal and 5 (19%) extra-adrenal in location. Patient characteristics are listed in (supplementary) Table 1. The institutional review board approved this retrospective study and the requirement to obtain informed consent was waived.

18F-FDG PET/CT ScanningAll patients underwent pre-surgical evaluation with 18F-FDG-PET/CT at the Radboud University Medical Center between December 2007 and February 2012 (Figure 1). Patients fasted for at least 6 hours before receiving a [mean±SD] 241±73 Mbq dose of intravenous 18F-FDG based on body weight. PET/CT scans were performed ~one hour (range; 55–74 minutes) after injection. Before November 2011, imaging was performed using a Biograph 2 PET/CT, and after November 2011 using an mCT-40 scanner (both Siemens Healthcare, USA). Both scanners were calibrated and certified by European Association of Nuclear Medicine (EANM) Research Ltd (EARL) in accordance with the EANM guidelines for PET/CT (25). First, a low-dose CT scan using “CareDose” with reference 40 mA and 130 kV was performed from the base of the skull to the mid-thigh. Instructions for breathing

89

18F-FDG PET in PPGL

5

and positioning were given to patients. The CT transaxial matrix size was 512 x 512 with pixels of 0.98 x 0.98 mm2 for both scanners. CT slice width was 3 mm for the Biograph 2 and 1.5 mm for the mCT. PET images were obtained using OSEM2D reconstruction on the Biograph 2 with 4 iterations and 16 subsets and a post reconstruction Gaussian filter of 5 mm FWHM. Transaxial PET matrix size was 128x128 and pixel size was 5.31 x 5.31 x 3 mm3. For the mCT, images were obtained using time-of-flight and HD reconstruction with 3 iterations and 21 subsets and a post reconstruction Gaussian filter of 8 mm FWHM. Transaxial PET matrix size was 256x256 and pixel size was 3.18 x 3.18 x 3 mm3. The large value of the Gaussian filter for the mCT images resulted in additional smoothing and was used in order to comply with the EANM guidelines for PET CT (25) allowing direct comparison of quantitative data from both scanners.

Image Interpretation and Quantitative MeasurementsPET/CT images were reviewed using Inveon Research Workplace version 4.1 software (Siemens Healthcare). Regions of interest (ROI) were manually drawn in each transversal slice over visually assessed lesions in correspondence with CT images. ROIs were combined to form a volume of interest (VOI), which was used for quantitative analysis. Maximum and mean standardized uptake values (SUVs) normalized for body weights were calculated as SUV = A / IA x BW [A: activity concentration of VOI (Bq/ml), BW: body weight (g), IA: injected activity (Bq)]. Liver-normalized SUVs were calculated as PPGL SUVs divided by corresponding liver mean SUVs in a fixed volume of interest in the upper central

Table 1 Patients characteristics

Genotype N (patients)

Gender Age, years

Tumour location Tumour dimensions,

cm3M F A EA

Sporadic* 15 8 7 54.2 ± 14.9 14 1 47.9 ± 56.3SDHB 2 2 0 31.0 ± 14.1 0 2 3.4 ± 2.4SDHD 3 2 1 49.5 ± 13.4 1 2 26.0 ± 33.8MEN-2 5 4 1 46.8 ± 14.4 5 0 14.8 ± 21.0

NF1 1 0 1 70 1 0 51.7MAX 1 1 0 63 1 0 20.4

*Presence of germline mutations and large deletions in SDHB/C/D, RET, VHL, and—since 2011—in SHDA, SDHAF2, TMEM127, and MAX was investigated using standard procedures, and no mutation was detected.Study included a total of 27 patients. Data are presented as mean ± SD.

90

liver. All calculated SUVs were decay corrected by using the following formula: A0 = At x eλt [A0: corrected activity, At = uncorrected activity, λ = decay constant (ln2/110) min-1, t = elapsed time in min].

Immunohistochemical Staining and Quantification and Quantitative Polymerase Chain Reaction (qPCR)See supplementary information for details.

Statistical Analysis Statistical analysis was conducted using SPSS 20 (SPSS Inc, Chicago, IL) and GraphPad Prism 6 software (Graphad Inc, La Jolla, CA). For comparison of immuno-histochemical staining scores and SUVs of different genotypes, scores and SUVs were analyzed using Kruskal Wallis test with post-hoc Dunn’s test. Results are presented as mean ± SD. Correlations were examined using Spearman correlation test. A two-side P value less than 0.05 was considered to be statistically significant.

Figure 1 Imaging results in a patient with a sporadic PPGL located in the left adrenal. (A) 18F-fluorodeoxyglucose positron emission tomography (B) 123I-metaiodobenzylguanidine single photon emission tomography (C) Computed tomography scan.

A B

C

91

18F-FDG PET in PPGL

5

5.3 Results

18F-FDG Uptake in PPGLs Maximum SUVs of PPGLs ranged from 0.8 to 11.8 (3.7 ± 3.2) and mean SUVs from 0.6 to 5.4 (1.8 ± 1.5). The distribution of SUVs in PPGLs across hereditary and sporadic tumours is shown in Figure 2. The maximum SUV for hereditary cluster 1 tumours (SDHB, SDHD) was higher (9.6 ± 1.5) than for hereditary cluster 2 tumours (RET, NF1, MAX: 1.9 ± 0.3, P<0.01) and sporadic tumours (2.6 ± 1.7, P<0.01). Also, mean SUVs for hereditary cluster 1 tumours were higher (4.5 ± 0.8) than for hereditary cluster 2 (1.0 ± 0.1, P<0.01) and sporadic tumours (1.3 ± 0.8, P<0.01). Maximum SUVs were higher for adrenal (2.7 ± 1.8, P<0.001) and extra-adrenal PPGLs (8.2 ± 4.3, P<0.01) than for normal adrenal glands (1.2 ± 0.5). Also, mean SUVs were higher for adrenal (1.3 ± 0.8, P<0.01) and extra-adrenal PPGL (4.0 ± 1.9, P<0.01) than for normal adrenal glands (0.8 ± 0.2). 18F-FDG uptake by normal adrenal glands did not exceed that of the liver in any case. No false-positive lesions were observed on 18F-FDG PET/CT. The volume of the PPGL lesions were calculated as described before (26) and varied from 0.8 to 161 cm3 (35 ± 46). There was no significant relation between tumour volume and 18F-FDG uptake (P=0.945).

Figure 2 18F-fluorodeoxyglucose positron emission tomography standardized uptake value (SUV) in PPGLs across different genotypes. (A) Maximum SUV (B) Mean SUV. All SUVs are normalized for body weight and liver and corrected for decay. **P < 0.01.

A B

92

Markers of Glucose Uptake and MetabolismAll 27 PPGL samples showed a positive cytoplasmic staining for HK-1. Staining was high across samples (Figure 3A) and no significant differences were observed between the different genotypes and sporadic samples (Figure 4). HK-2 staining showed a clear variation between samples (Figure 3A). Negative to occasionally medium cytoplasmic staining was encountered in sporadic tumour samples. HK-2 expression was found to be significantly higher in SDHx-related PPGLs than in sporadic (P=0.022) and RET, NF-1 and MAX-related tumours (P=0.039) (Figure 4). Increased expression of HK-2 was confirmed at the mRNA level using qPCR and a high (P=0.004) expression of the markers was observed in SDHx-related tumours when compared to sporadic tumours and cluster 2-related tumours (supplementary Figure 1). HK-3 expression was significantly higher in SDHx-related PPGLs than in sporadic tumours (P=0.025),

Figure 3 Immunohistochemical staining of PPGLs for hexokinases. Representative images of a SDHB, RET and sporadic tumour with magnifications as indicated. Areas with brown color (DAB polymer) are representative of positive staining.

93

18F-FDG PET in PPGL

5

however when compared to RET, NF-1 and MAX-related tumours no statistical significance was observed (Figure 4). GLUT-1 and GLUT-3 showed predominantly cytoplasmic staining and occasionally (5-10%) cell membrane staining (Figure 3B). Sporadic, SDHx, RET, NF-1 and MAX-related tumours showed an overall similar GLUT-1 expression (Figure 5). A homogeneous distribution of staining in the cytoplasm was observed. Sporadic and RET associated tumours showed a similar GLUT-3 expression, which was usually scored as medium. SDHx-related PPGLs appeared to exhibit a higher GLUT-3 staining compared to sporadic and RET PPGLs. However, statistical significance was not achieved (Figure 6). On the other hand, at the mRNA level, GLUT-3 expression did not show significant differences between the groups (supplementary Figure 1). MCT-4 was localized to cytoplasm as well as membrane.

Figure 4 Comparison of staining for HK-1 (A), HK-2 (B), and HK-3 (C) among different PPGL genotypes. Graphs represent staining scores, calculated as percentage area stained positive times staining intensity. Groups with different letters as superscripts are significantly different.

A

B

C

94

Figu

re 5

Imm

unoh

isto

chem

ical

stai

ning

of P

PGLs

for G

LUT-

1 and

GLU

T-3,

MCT

-4, a

nd V

EGF.

Repr

esen

tati

ve im

ages

of a

SDH

B, R

ET a

nd sp

orad

ic

tum

our w

ith

mag

nific

atio

ns a

s ind

icat

ed. A

reas

wit

h br

own

colo

r (D

AB

poly

mer

) are

repr

esen

tati

ve o

f pos

itiv

e st

aini

ng.

95

18F-FDG PET in PPGL

5

The expression of MCT-4 did not show significant differences between the different genotypes (Figures 3B and 6). VEGF expression was found to be significantly higher in SDHx-related PPGLs compared to RET, NF-1 and MAX-related PPGLs (P=0.008) (Figure 3C and 6). There was a significant difference in mean percentage anti-CD34 stained area (P=0.050) and mean vessel perimeter (P=0.010) between SDHx-related and sporadic PPGLs. The microvessel density did not differ among genotypes.

Relationship between 18F-FDG Uptake and Markers of Glucose Uptake and MetabolismThe correlative relationships between immunohistochemical staining and calculated SUVs and corresponding P-values are summarized in Table 2. Mean SUVs of PPGLs were significantly associated with HK-2, HK-3, VEGF and MCT-4 staining. The strongest correlation with SUV was found for HK-3 expression (R=0.471, P=0.013). No significant relationship was found between GLUT-1 and GLUT-3 staining and SUV. Maximum SUVs showed only a significantly correlative relationship for HK-2 (R=0.409, P=0.034) and HK-3 immunohistochemical staining

Figure 6 Comparison of staining for GLUT-1 (A), GLUT-3 (B), VEGF (C), and MCT-4 (D) among different PPGL genotypes. Graphs represent staining scores, calculated as percentage area stained positive times staining intensity. Groups with different letters as superscripts are significantly different.

A

C

B

D

96

(R=0.381, P=0.050). No correlative relationships were found for 18F-FDG uptake and vessel parameters.

5.4 Discussion

We examined the ex vivo expression of markers of the Warburg effect in PPGLs using immunohistochemical staining and their correlation with uptake of in vivo 18F-FDG on PET/CT scans. We confirmed genotype-specific differences in 18F-FDG uptake where SDHx-related PPGLs showed highest SUVs. The expression of HK-2 was significantly higher in SDHx-related PPGLs compared to sporadic and cluster 2 PPGLs. Furthermore, expression of HK-3 was significantly higher in SDHx-related PPGLs compared to sporadic PPGLs and expression of VEGF was significantly higher in SDHx-related PPGLs compared to cluster 2 PPGLs. Uptake of 18F-FDG significantly correlated with the expression of HK-2, HK-3, VEGF and MCT-4. Similar to glucose, 18F-FDG is taken up by tumour cells mostly via facilitative transport by GLUTs. After cell entry, 18F-FDG is phosphorylated by HKs into 18F-FDG-6-P which, in contrast to glucose-6-P, cannot be further metabolized along the glycolytic pathway. The cell membrane is impermeable to 18F-FDG-6P so it accumulates within cells directly proportionate to their metabolic activity. 18F-FDG-6P can theoretically escape from the cell by dephosphorylation back to

Table 2 Correlations between 18F-fluorodeoxyglucose uptake and immuno-histochemical markers of glucose uptake and metabolism

Marker SUV Maximum SUV MeanR P R P

HK-2 0.409 (0.034)* 0.425 (0.027)*

HK-3 0.381 (0.050)* 0.471 (0.013)*

GLUT-1 -0.001 (0.998) -0.100 (0.620)

GLUT-3 0.174 (0.386) 0.263 (0.185)

VEGF 0.312 (0.114) 0.383 (0.049)*

MCT-4 0.373 (0.055) 0,444 (0.020)*

All SUVs are normalized to liver normalized and decay corrected. Correlations are expressed as Spearman’s rho and P-value (* = P < 0.05).

97

18F-FDG PET in PPGL

5

18F-FDG by glucose-6-phosphatase. In general, this process is negligible because of the low intracellular levels of the dephosphorylating enzyme. Therefore, 18F-FDG uptake of any cell is determined by expression of GLUTs and activity of HKs. Another possible determinant is tumoural blood flow which brings 18F-FDG to the cell and reasonably increases its metabolism in parallel. However, activation of the hypoxic-angiogenic pathway via HIF-1α and -2α results in adaptation of the tumour to a hypoxic environment and consequently an uncoupling of blood flow and metabolism implying hypoxic stimulation of glucose metabolism in the presence of adequate oxygen. This phenomenon, known as the Warburg effect (27), characterizes cluster 1 PPGLs (VHL, SDHx). Favier et al. demonstrated that HIF-2α, which together with HIF-1α upregulates VEGF and GLUT-1 gene expression (28), is overexpressed in VHL and SDHx-related PPGLs (29). VEGF activated in endothelial vascular cells within tumours can significantly contribute to 18F-FDG uptake (30). High SUVs observed in SDHx-related PPGLs could be related to the Warburg effect and thus we expected to find a stronger GLUT-1 and GLUT-3 expression in SDHx- compared to cluster 2-related PPGLs. However, we found no significant increase in immunohistochemical GLUT staining in SDHx-related tumours. We observed a predominantly cytoplasmic and weak membrane staining for both GLUT-1 and GLUT-3 as also reported by Blank et al (31). In addition, we found no correlation between GLUTs and SUVs in PPGLs. Taken together, these results suggest that glucose uptake may not be preferentially regulated by GLUT-1 and GLUT-3 in these tumours. Aloj et al. concluded that higher levels of glucose transporter protein do not guarantee increased FDG uptake by cancer cells (32). Differences in 18F-FDG uptake may be caused by differences in GLUT transporter activity rather than by differences in number of transporters present. Also, our findings suggest that the expression and recruitment of GLUTs to the cell membrane do not appear to vary in a genotype-specific way. Instead of GLUT, higher 18F-FDG uptake in SDHx PPGLs might be related to expression of HKs. HK-2 showed an increased expression in SDHx-related PPGLs as predicted. The expression of HK-2 is regulated by HIF-1 and it is predominantly overexpressed in various cancer cells that display the Warburg effect (18) and associated with 18F-FDG uptake (20, 33). Favier et al. (10) also observed a significantly higher HK-2 gene expression in SDHx-related PPGLs compared to NF1 and RET related PPGLs. Immunohistochemical staining of HK-3 showed increased expression in SDHx-related PPGLs compared to sporadic tumours but when compared with cluster 2 tumours no significant increased expression was found. In addition, we found a stronger correlation between HK-3 and SUV than HK-2. A study using rat hepatoma cell line N1S1 reported the role of hypoxia signaling in the regulation of HK-3 (22). However, little is known about its regulation by HIFs

98

in PPGLs. Thus, increased glucose phosphorylation by HKs rather than the glucose transport by GLUTs is reflected as increased 18F-FDG uptake. Also, we assessed expression of HK-1 and observed lack of differences among different genotypes as it is a housekeeping enzyme ubiquitously expressed in mammalian tissues independent of HIF and is unaltered in most cancer cells (34). Increase in HK-2 and HK-3 expression in SDHx-related PPGLs point towards increased glycolysis, which we further investigated by assessing expression of MCT-4. This monocarboxylate transporter exports lactate out of the cell and its expression is known to be high in cells with increased glycolysis and lactate production. We found membrane as well as cytoplasmic staining for MCT-4 but the levels of expression did not vary among the different genotypes. However, Favier et al. (10) reported an increased expression of MCT-4 at the mRNA level in SDHx-related tumours. Thus, though evidence suggests increased glycolysis in SDHx-related tumours, it might not be accompanied by increased lactate production. We determined expression of VEGF to further explore our hypothesis that higher 18F-FDG uptake in SDHx-related PPGLs is reflective of the Warburg effect. A significantly increased expression of VEGF was observed, suggesting an increase in microvessel density in SDHx-related PPGLs. These results are in line with our previous estimations of mRNA levels measured by qPCR (35) and with Blank et al. (31) who also observed a significant correlation between SDHB-related PPGLs and microvessel density. VEGF stimulates angiogenesis but this probably will not necessarily result in higher uptake of 18F-FDG since newly formed vessels can be leaky and do not provide tumour cells with nutrients like glucose. Therefore, we also investigated CD34, marker for more mature endothelial cells. An increase in percentage of anti-CD34 staining was confirmed by quantitative analysis of CD34 immunohistochemical staining. Moreover, mean vascular perimeter was significantly higher in SDHx-related tumours compared to cluster 2 and sporadic tumours. Nevertheless, no statistical differences in microvessel density were observed which could be attributed to highly heterogenous vascular patterns observed in these tumours. Taken together, increased endothelial surface area in this type of tumours can potentially contribute to higher accumulation of 18F-FDG, supporting the idea that functionality of vessels is more critical for SUVs than microvessel density. Additional studies with a larger sample size are needed to further explore the genotype-specific signature of expression of markers of the Warburg effect and their impact on functional imaging of these tumours. Also, somatic mutations in apparently sporadic tumours will need to be taken into account. Besides genetic factors, the influence of the microenvironment on the metabolism of tumour cells deserves further investigations. Furthermore, besides immunohistochemistry,

99

18F-FDG PET in PPGL

5

a more quantitative approach such as western blotting (not performed in the present study due to limited sample quantity) would be useful for better assessment of the expression of the investigated markers. In addition, SUVs merely provide a semi-quantitative measurement of 18F-FDG uptake, not providing information on the actual exchange of 18F-FDG between the intra- and extracellular compartment. This would require a dynamic scanning approach. In addition, considering the heterogeneous uptake pattern observed in some tumours, both SUVmean and SUVmax have their limitations regarding their representativeness for the tumour as a whole and the correlation with tissue markers.

5.5 Conclusion

Activation of aerobic glycolysis in SDHx-related PPGLs is associated with increased 18F-FDG accumulation due to accelerated glucose phosphorylation by hexokinases rather than increased expression of glucose transporters. Differences in tumour vasculature and the activity of transporter systems may also contribute to genotype-related SUVs.

100

5.6 References1. Gimenez-Roqueplo AP, Dahia PL, Robledo M. An update on the genetics of paraganglioma, pheochro-

mocytoma, and associated hereditary syndromes. Horm Metab Res. 2012;44(5):328-333.2. Eisenhofer G, Huynh TT, Pacak K, Brouwers FM, Walther MM, Linehan WM, Munson PJ, Mannelli M,

Goldstein DS, Elkahloun AG. Distinct gene expression profiles in norepinephrine- and epineph-rine-producing hereditary and sporadic pheochromocytomas: activation of hypoxia-driven angiogenic pathways in von Hippel-Lindau syndrome. Endocr Relat Cancer. 2004;11(4):897-911.

3. Lopez-Jimenez E, Gomez-Lopez G, Leandro-Garcia LJ, Munoz I, Schiavi F, Montero-Conde C, de Cubas AA, Ramires R, Landa I, Leskela S, Maliszewska A, Inglada-Perez L, de la Vega L, Rodriguez-Antona C, Leton R, Bernal C, de Campos JM, Diez-Tascon C, Fraga MF, Boullosa C, Pisano DG, Opocher G, Robledo M, Cascon A. Research resource: Transcriptional profiling reveals different pseudohypoxic signatures in SDHB and VHL-related pheochromocytomas. Mol Endocrinol. 2010;24(12):2382-2391.

4. Timmers HJ, Chen CC, Carrasquillo JA, Whatley M, Ling A, Havekes B, Eisenhofer G, Martiniova L, Adams KT, Pacak K. Comparison of 18F-fluoro-L-DOPA, 18F-fluoro-deoxyglucose, and 18F-fluorodopa-mine PET and 123I-MIBG scintigraphy in the localization of pheochromocytoma and paraganglioma. The Journal of clinical endocrinology and metabolism. 2009;94(12):4757-4767.

5. Timmers HJ, Chen CC, Carrasquillo JA, Whatley M, Ling A, Eisenhofer G, King KS, Rao JU, Wesley RA, Adams KT, Pacak K. Staging and functional characterization of pheochromocytoma and paraganglioma by 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography. Journal of the National Cancer Institute. 2012;104(9):700-708.

6. Timmers HJ, Kozupa A, Chen CC, Carrasquillo JA, Ling A, Eisenhofer G, Adams KT, Solis D, Lenders JW, Pacak K. Superiority of fluorodeoxyglucose positron emission tomography to other functional imaging techniques in the evaluation of metastatic SDHB-associated pheochromocytoma and paraganglioma. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2007;25(16):2262-2269.

7. Taieb D, Sebag F, Barlier A, Tessonnier L, Palazzo FF, Morange I, Niccoli-Sire P, Fakhry N, De Micco C, Cammilleri S, Enjalbert A, Henry JF, Mundler O. 18F-FDG avidity of pheochromocytomas and paragangliomas: a new molecular imaging signature? Journal of nuclear medicine : official publication, Society of Nuclear Medicine. 2009;50(5):711-717.

8. Blanchet EM, Gabriel S, Martucci V, Fakhry N, Chen CC, Deveze A, Millo C, Barlier A, Pertuit M, Loundou A, Pacak K, Taieb D. F-FDG PET/CT as a predictor of hereditary head and neck paragangliomas. European journal of clinical investigation. 2014.

9. Gimenez-Roqueplo AP, Favier J, Rustin P, Rieubland C, Kerlan V, Plouin PF, Rotig A, Jeunemaitre X. Functional consequences of a SDHB gene mutation in an apparently sporadic pheochromocytoma. The Journal of clinical endocrinology and metabolism. 2002;87(10):4771-4774.

10. Favier J, Briere JJ, Burnichon N, Riviere J, Vescovo L, Benit P, Giscos-Douriez I, De Reynies A, Bertherat J, Badoual C, Tissier F, Amar L, Libe R, Plouin PF, Jeunemaitre X, Rustin P, Gimenez-Roqueplo AP. The Warburg effect is genetically determined in inherited pheochromocytomas. PLoS One. 2009;4(9):e7094.

11. Rao JU, Engelke UF, Rodenburg RJ, Wevers RA, Pacak K, Eisenhofer G, Qin N, Kusters B, Goudswaard AG, Lenders JW, Hermus AR, Mensenkamp AR, Kunst HP, Sweep FC, Timmers HJ. Genotype-specific abnormalities in mitochondrial function associate with distinct profiles of energy metabolism and catecholamine content in pheochromocytoma and paraganglioma. Clinical cancer research : an official journal of the American Association for Cancer Research. 2013;19(14):3787-3795.

12. Jochmanova I, Yang C, Zhuang Z, Pacak K. Hypoxia-inducible factor signaling in pheochromocyto-ma: turning the rudder in the right direction. J Natl Cancer Inst. 2013;105(17):1270-1283.

13. Gordan JD, Simon MC. Hypoxia-inducible factors: central regulators of the tumour phenotype. Current opinion in genetics & development. 2007;17(1):71-77.

14. Dahia PL, Familial Pheochromocytoma C. Transcription association of VHL and SDH mutations link hypoxia and oxidoreductase signals in pheochromocytomas. Annals of the New York Academy of Sciences. 2006;1073:208-220.

101

18F-FDG PET in PPGL

5

15. Selak MA, Armour SM, MacKenzie ED, Boulahbel H, Watson DG, Mansfield KD, Pan Y, Simon MC, Thompson CB, Gottlieb E. Succinate links TCA cycle dysfunction to oncogenesis by inhibiting HIF-alpha prolyl hydroxylase. Cancer cell. 2005;7(1):77-85.

16. Warburg O. On the origin of cancer cells. Science. 1956;123(3191):309-314.17. Macheda ML, Rogers S, Best JD. Molecular and cellular regulation of glucose transporter (GLUT)

proteins in cancer. Journal of cellular physiology. 2005;202(3):654-662.18. Mathupala SP, Ko YH, Pedersen PL. Hexokinase-2 bound to mitochondria: cancer’s stygian link to the

“Warburg Effect” and a pivotal target for effective therapy. Seminars in cancer biology. 2009;19(1):17-24.19. Ahn KJ, Hwang HS, Park JH, Bang SH, Kang WJ, Yun M, Lee JD. Evaluation of the role of hexokinase

type II in cellular proliferation and apoptosis using human hepatocellular carcinoma cell lines. Journal of nuclear medicine : official publication, Society of Nuclear Medicine. 2009;50(9):1525-1532.

20. Yamada T, Uchida M, Kwang-Lee K, Kitamura N, Yoshimura T, Sasabe E, Yamamoto T. Correlation of metabolism/hypoxia markers and fluorodeoxyglucose uptake in oral squamous cell carcinomas. Oral surgery, oral medicine, oral pathology and oral radiology. 2012;113(4):464-471.

21. de Geus-Oei LF, van Krieken JH, Aliredjo RP, Krabbe PF, Frielink C, Verhagen AF, Boerman OC, Oyen WJ. Biological correlates of FDG uptake in non-small cell lung cancer. Lung cancer. 2007;55(1):79-87.

22. Wyatt E, Wu R, Rabeh W, Park HW, Ghanefar M, Ardehali H. Regulation and cytoprotective role of hexokinase III. PloS one. 2010;5(11):e13823.

23. Dierckx RA, Van de Wiele C. FDG uptake, a surrogate of tumour hypoxia? European journal of nuclear medicine and molecular imaging. 2008;35(8):1544-1549.

24. Clavo AC, Brown RS, Wahl RL. Fluorodeoxyglucose uptake in human cancer cell lines is increased by hypoxia. Journal of nuclear medicine : official publication, Society of Nuclear Medicine. 1995;36(9):1625-1632.

25. Boellaard R, O’Doherty MJ, Weber WA, Mottaghy FM, Lonsdale MN, Stroobants SG, Oyen WJ, Kotzerke J, Hoekstra OS, Pruim J, Marsden PK, Tatsch K, Hoekstra CJ, Visser EP, Arends B, Verzijlbergen FJ, Zijlstra JM, Comans EF, Lammertsma AA, Paans AM, Willemsen AT, Beyer T, Bockisch A, Schae-fer-Prokop C, Delbeke D, Baum RP, Chiti A, Krause BJ. FDG PET and PET/CT: EANM procedure guidelines for tumour PET imaging: version 1.0. European journal of nuclear medicine and molecular imaging. 2010;37(1):181-200.

26. Eisenhofer G, Pacak K, Huynh TT, Qin N, Bratslavsky G, Linehan WM, Mannelli M, Friberg P, Grebe SK, Timmers HJ, Bornstein SR, Lenders JW. Catecholamine metabolomic and secretory phenotypes in phaeochromocytoma. Endocrine-related cancer. 2011;18(1):97-111.

27. Vander Heiden MG, Cantley LC, Thompson CB. Understanding the Warburg effect: the metabolic requirements of cell proliferation. Science. 2009;324(5930):1029-1033.

28. Semenza GL. Hypoxia-inducible factors in physiology and medicine. Cell. 2012;148(3):399-408.29. Favier J BJ, Burnichon N, Rivière J, Vescovo L, Benit P, Giscos-Douriez I, De Reyniès A, Bertherat J,

Badoual C, Tissier F, Amar L, Libé R, Plouin PF, Jeunemaitre X, Rustin P, Gimenez-Roqueplo AP. The Warburg effect is genetically determined in inherited pheochromocytomas. PLoS One. 2009;18(4(9)).

30. Maschauer S PO, Hoffmann M, Deichen JT, Kuwert T. Characterization of 18F-FDG uptake in human endothelial cells in vitro. J Nucl Med. 2004;45(3):455-460.

31. Blank A, Schmitt AM, Korpershoek E, van Nederveen F, Rudolph T, Weber N, Strebel RT, de Krijger R, Komminoth P, Perren A. SDHB loss predicts malignancy in pheochromocytomas/sympathethic paragangliomas, but not through hypoxia signalling. Endocrine-related cancer. 2010;17(4):919-928.

32. Aloj L, Caraco C, Jagoda E, Eckelman WC, Neumann RD. Glut-1 and hexokinase expression: relationship with 2-fluoro-2-deoxy-D-glucose uptake in A431 and T47D cells in culture. Cancer Res. 1999;59(18):4709-4714.

33. Watanabe Y, Suefuji H, Hirose Y, Kaida H, Suzuki G, Uozumi J, Ogo E, Miura M, Takasu K, Miyazaki K, Nakahara K, Ishibashi M, Okamura T, Ohshima K, Hayabuchi N. 18F-FDG uptake in primary gastric malignant lymphoma correlates with glucose transporter 1 expression and histologic malignant potential. International journal of hematology. 2013;97(1):43-49.

34. Wolf A, Agnihotri S, Micallef J, Mukherjee J, Sabha N, Cairns R, Hawkins C, Guha A. Hexokinase 2 is a key mediator of aerobic glycolysis and promotes tumour growth in human glioblastoma multiforme. The Journal of experimental medicine. 2011;208(2):313-326.

102

35. Span PN, Rao JU, Oude Ophuis SB, Lenders JW, Sweep FC, Wesseling P, Kusters B, van Nederveen FH, de Krijger RR, Hermus AR, Timmers HJ. Overexpression of the natural antisense hypoxia-inducible factor-1alpha transcript is associated with malignant pheochromocytoma/paraganglioma. Endocrine- related cancer. 2011;18(3):323-331.

36. Vlems F, van der Worp E, van der Laak J, van de Velde C, Nagtegaal I, van Krieken H. A study into methodology and application of quantification of tumour vasculature in rectal cancer. Virchows Archiv : an international journal of pathology. 2004;445(3):263-270.

103

18F-FDG PET in PPGL

5

5.7 Supplementary Methods

Quantitative PCRTotal RNA was isolated using the Total RNA purification Kit (Norgen Biotek Corp) with on-column DNase-I treatment. RNA concentrations were determined by measuring the spectrophotometric absorption at 260 nm using the GeneQuant (Amersham). cDNA synthesis was performed using 500 ng total RNA and 8 µl 5X iScript reaction mix (Bio-Rad) containing oligo dT primers, with a total volume of 40 µl. cDNA synthesis was performed using a heating block. Samples were denatured for 10 min at 700C and cDNA synthesis was performed for 60 min at 420C, followed by an enzyme inactivation step for 5 min at 990C. Quantitative PCR (qPCR) was performed to determine the expression levels of HK-2, GLUT-3 and β2 Microglobulin (β2M) mRNAs in 22 PPGL tumour tissues. The primers used for the reactions are as follows: HK-2 Forward Primer 5’ TCAGATTGAGAGTGACTGCC 3’ and Reverse Primer 5’ TTTCTCGTATCCTGTCCACC 3’ GLUT-3 Forward Primer 5’GTG-GCCCAGATCTTTGGTC3’ and Reverse Primer 5’ AAGGGCTGCACTTTGTAGG3’ β2M Forward Primer 5’ATGAGTATGCCTGCCGTGTG3’ Reverse Primer 5’CCAAATGCGG-CATCTTCAAAC3’. Reactions were performed using SYBR green PCR master mix (Applied Biosystems). cDNA equivalent of 10 ng total RNA was used for each reaction and the reactions were performed in a total volume of 10 µl in ABI Prism 7500 sequence detector (Applied Biosystems). The amplification reactions were carried out with denaturation at 95 0C for 10 min, 40 cycles of 15 s at 950C (melting), and 60 s at 600C (annealing and elongation). The expression of the mRNAs was calculated relative to the levels of the internal control gene β2M. The relative expression levels were determined by the 2-ΔCt method.

Immunohistochemical staining and quantificationTumour tissue sections were deparaffinized, rehydrated and washed with 50 mM phosphate buffered saline (PBS). Antigen retrieval was performed by boiling the sections in 10 mM citrate, pH 6.0 in a microwave for 3 minutes at 850W followed by 10 minutes at 150 W. Endogenous peroxidase activity was blocked with 3% hydrogen peroxide (H2O2) in PBS for 30 minutes. For HK-1, 2 and 3 staining, endogenous peroxidase activity was blocked by incubation with 0.6% H2O2 in 40% methanol and 60% 50 mM PBS for 30 minutes. Subsequently, the sections were washed with PBS. Non-specific interactions were blocked using normal human serum (5% in PBS for 30 minutes) for HK-1, 2 and 3, normal goat serum (20% in PBS) for VEGF, MCT-4, (30 minutes) GLUT-1 and -3 (10 minutes) respectively. For GLUT-1 prior to incubation with the primary antibody, avidin and biotin block (Vector Laboratories, SP-2001) was performed for 15 minutes each. Primary antibody diluted in 0.1% BSA/PBS was applied (supplementary Table 2) for 60

104

minutes for HK-1, HK-2, HK-3, CD34 and GLUT-3, 90 minutes for MCT-4 and overnight for GLUT-1 and VEGF at 4 °C. The sections were then washed with 50 mM PBS, followed by incubation with the secondary antibody for 30 minutes (supplementary Table 2). Slides stained for HK-1, HK-2, HK-3 and GLUT-1 were incubated with avidin biotin reagent (Vectastain Elite ABC Kit, Vector Laboratories, PK6100) for 45 minutes. The sections were washed with PBS and incubated with 3,3-diamino-benzidine (DAB) (Immunologic, BS04-110) for 10 minutes. After washing DAB with running tap water for 5 minutes, the sections were counterstained with hematoxylin for 10 seconds, washed and dehydrated in ethanol series (50%, 70%, 100%) followed by xylene and mounted with permount. Expression of various markers was quantified by evaluating the amount of DAB staining by an experienced pathologist (BK) who was unaware of the patient’s clinical data while CD34 was scored automatically. The percentage of the positive stained area of the sections was graded in a scale ranging from 0-100%. The intensity of the DAB color was graded as non-staining (0), faint staining (I), medium staining (II) and intense staining (III), ignoring the background color of the staining. Areas containing adrenal cortex, brown fat and capillaries were used as controls for comparison between different sections. The intensity of each marker was evaluated separately. Cytoplasmic staining was evaluated for HK-1, HK-2, HK-3 and VEGF. For GLUT-1, GLUT-3 and MCT-4 cytoplasmic and/or membrane staining was graded as they were expressed both in the cytoplasm and membrane. The score of the staining represents the expression of the markers and is based on the multiplication of the intensity with the percentage of the positive stained area. For CD34 scoring, image acquisition was performed using a CCD color camera (AxioCam MRc; Carl Zeiss AG, Jena, Germany) mounted on a conventional light microscope (AxioPhot; Carl Zeiss AG, Jena, Germany) and attached to a personal computer. Images were acquired using a x20 objective (Plan Neofluar, NA = 0.5, resulting in specimen level pixel size of 0.53x0.53 µm2). Prior to analysis of the immunohistochemical staining, an image of an empty microscopic field was acquired, which was used for correction for unequal illumination. Image acquisition and analysis were performed using a custom written macro in KS400 image analysis software (Carl Zeiss), as described previously (36). Thresholds for recognition of CD34-positivity were determined from a set of training slides and were found adequate for almost all slides analyzed in this study. When the initial thresholds led to unrealistic patterns, interactive adjustment was performed by the operator. For each patient, the percentage of anti-CD34 stained area, the microvessel density and the vessel perimeter were calculated per surface area tumour in ten randomly selected images.

105

18F-FDG PET in PPGL

5

Supp

lem

enta

ry T

able

1

Gen

otyp

eSe

xA

ge

(y)

Tum

our

dim

ensi

ons

(mm

)

Tum

our

loca

tion

Tum

our s

tage

HK-2

HK-3

GLUT-1

GLUT-3

VEGF

MCT-4

SUV max

SUV mean

Spor

adic

1F

7012

x 8

x 16

EA

met

asta

sis

200

100

120

100

200

900.

790.

70Sp

orad

icF

6630

x 2

7 x

32RA

prim

ary

8080

9090

905

2.78

1.27

Spor

adic

F56

21 x

17 x

22

RApr

imar

y 40

8011

013

016

010

01.6

70.

78Sp

orad

icM

6627

x 2

3 x

23RA

rest

tum

our

060

080

9090

1.93

1.14

Spor

adic

F75

35 x

26

x 37

RApr

imar

y27

080

2518

027

010

05.

822.

81Sp

orad

icM

4423

x 2

6 x

32RA

prim

ary

9060

4015

080

602.

321.0

8Sp

orad

icM

7052

x 4

0 x

50RA

pr

imar

y 0

4020

05

201.2

80.

62Sp

orad

icM

4122

x 2

2 x

21RA

prim

ary

190

120

6621

018

030

0.83

0.60

Spor

adic

M36

70 x

63

x 60

RApr

imar

y 20

6010

010

150

402.

150.

90Sp

orad

icM

4975

x 5

6 x

73RA

prim

ary

010

165

110

9016

02.

280.

97Sp

orad

icF

5335

x 3

2 x

33LA

prim

ary

5012

012

015

016

016

01.8

71.1

0Sp

orad

icM

2078

x 5

2 x

70RA

prim

ary

160

2015

180

180

270

4.58

1.63

Spor

adic

M62

31 x

35

x 38

LApr

imar

y 18

030

160

9018

016

01.6

40.

95Sp

orad

icF

5050

x 5

2 x

53LA

prim

ary

9016

060

150

150

270

3.06

1.62

Spor

adic

F55

43 x

40

x 50

LApr

imar

y 11

016

020

090

140

160

6.47

3.26

RET

M36

17 x

13 x

20

LAlo

cal r

elap

se 2

3516

030

160

9080

1.61

1.01

RET

M36

23 x

19 x

30

LApr

imar

y 15

040

8515

018

040

1.57

0.83

RET

M44

27 x

18 x

38

RApr

imar

y 20

180

100

160

130

200

2.42

1.25

106

Supp

lem

enta

ry T

able

1 C

onti

nued

Gen

otyp

eSe

xA

ge

(y)

Tum

our

dim

ensi

ons

(mm

)

Tum

our

loca

tion

Tum

our s

tage

HK-2

HK-3

GLUT-1

GLUT-3

VEGF

MCT-4

SUV max

SUV mean

RET

M71

22 x

17 x

17RA

prim

ary

5016

050

160

9016

01.8

30.

99RE

TF

4749

x 4

5 x

45RA

prim

ary

3070

1050

150

701.9

01.1

1M

AX

M63

38 x

25

x 41

LAlo

cal r

elap

se 1

100

9030

7090

302.

300.

99N

F1F

7046

x 3

3 x

65RA

pr

imar

y 70

7090

7010

701.9

70.

94SD

HD

M32

21 x

14 x

21

EApr

imar

y12

018

030

200

200

160

10.2

3.99

SDH

DF

6055

x 4

5 x

50EA

pr

imar

y 21

014

011

020

027

060

9.28

5.42

SDH

DM

4630

x 3

0 x

21RA

prim

ary

280

160

6090

180

801.6

40.

95SD

HB

M41

22 x

22

x 20

EA

loca

l rel

apse

223

020

020

090

180

160

11.8

4.93

SDH

B M

2113

x 14

x 17

EA

prim

ary

200

9020

150

270

270

8.83

4.76

Abb

revi

aton

s: M

=mal

e; F

=fem

ale;

y =

yea

rs; E

A =

ext

ra-a

dren

al; L

A =

left

adr

enal

; RA

= ri

ght a

dren

al; H

K-2

= he

xoki

nase

-2; H

K-3

= he

xoki

nase

-3; G

LUT-

1 = g

luco

se

tran

spor

ter t

ype

1; G

LUT-

3 =

gluc

ose

tran

spor

ter t

ype

3; V

EGF

= va

scul

ar e

ndot

helia

l gro

wth

fact

or; M

CT-4

= m

onoc

arbo

xyla

te tr

ansp

orte

r typ

e 4;

SU

V =

stan

dard

up

take

val

ue (m

ax =

max

imum

).

1 Pat

ient

s wer

e te

sted

for t

he p

rese

nce

of g

erm

line

mut

atio

ns a

nd la

rge

dele

tion

s in

SDH

B/C/

D, R

ET, V

HL a

nd -s

ince

201

1- in

SDH

A, S

DH

AF2

, TM

EM12

7 and

MA

X.

107

18F-FDG PET in PPGL

5

Supplementary Table 2 Details of antibodies used for immunohistochemistry

Primary antibody

Host Concentration (µg/ml)

Manufacturer / Catalog number

Secondary antibody

Concentration (µg/ml)

Manufacturer / Catalog number

HK-1 Goat 10 Santa Cruz Biotechnology

SC 6518

Rabbit anti goat biotin

4 DAKO, EO466

HK-2 Goat 3 Santa Cruz Biotechnology

SC 6521HK-3 Goat 10 Santa Cruz

Biotechnology SC 6523

GLUT-1 Rabbit 1 Thermo Scientific RB-

9052-PO

Goat anti rabbit biotin

7.5 Vector Laboratories,

BA-1000GLUT-3 Mouse Neat Thermo

Scientific PA1-28204

Goat polyclonal

anti- mouse/

rat/ rabbit peroxidase

1:1 of neat Immunologic, DPVO 110HRP

MCT-4 Rabbit 5 Millipore AB3316P

VEGF Rabbit 32 Gifted by dept. of Laboratory

Medicine, RUNMC

CD-34 Mouse 1:200 Immunologic ILM 1343

Immunologic, DVPO 999HRP

108

Supplementary Figure 1 Expression of (A) Hexokinase 2 and (B) glucose transporter type 3 mRNA in PGL tumours. Expression of mRNAs in PPGL tumours were investigated using qPCR and their relative expression levels across different genotypes are plotted. Groups with different letters as superscripts are significantly different.

A

B

109

18F-FDG PET in PPGL

5

Discussion

6

113

Discussion

6

The focus of this doctoral thesis is to investigate the impact of genotype-specific differences in energy metabolism of PPGLs on catecholamine metabolism, imaging and metastatic potential. Initial work in this area focused on understanding differences in gene expression between cluster 1 and cluster 2 PPGLs. Cluster 1 tumours (VHL, SDHA/B/C/D/AF2) are characterized by increased expression of genes involved in (pseudo) hypoxia, cell proliferation, angiogenesis, electron transport chain and the Krebs cycle, and abnormal function of oxidore-ductases (1). Cluster 2 tumours (RET, NF1) show an increased expression of genes involved in protein synthesis, kinase signaling, endocytosis, and maintenance of a differentiated chromaffin cell catecholamine biosynthetic and secretory phenotype. Subsequent studies focused on alterations in respiratory chain enzyme activities in SDHx tumours in comparison with tumours of other genotypes (2,3). Studies on catecholamine metabolism and secretory phenotype also investigated the causes for the observed genotype-specific differences in catecholamine biosynthetic and secretory profile (4,5). In parallel, studies also investigated the efficacy of 18F-FDG-PET imaging in localizing metastatic PPGLs especially those with an underlying SDHB mutation (6,7). How genotype-specific alterations in energy metabolism underlie these various pathophysiological aspects of PPGLs was investigated in a series of studies composing the present thesis.

6.1 Relating energy metabolism to catecholamine, amino acid and purine metabolism in PPGLs

Towards this end, in the study that constitutes the first chapter of this thesis, we investigated genotype-specific differences in mitochondrial energy metabolism in PPGLs. The study provided novel insight into how deregulation of energy metabolism might impact catecholamine phenotypic features of cluster 1 and cluster 2 tumours. In accordance with previous reports (2,3,8,9), we observed that the activity of SDH (complex II) is low in SDH-related tumours. Interestingly, reduction of complex II activity in SDH-related tumours was associated with increased activities of respiratory chain complexes I, III, and IV and citrate synthase. The increased citrate synthase activity also indicated an increase in the mitochondrial content in SDH-related tumours. A similar increase in activities of complexes I and III and citrate synthase in SDH-related tumours when compared to VHL-related ones was also observed by Fliedner et al. (10). However, despite the apparent increase in activity of complexes I, III, and IV and citrate synthase in the tumours, ATP/ADP/AMP levels were low. Interestingly, in contrast to SDH-related tumours, we observed a trend of down regulation of all respiratory chain enzyme

114

activities in VHL tumours. Also, we observed that in contrast to VHL tumours, levels of ATP/ADP/AMP in SDHx tumours were undetectable. Genotype-specific differences in catecholamine production in PPGLs are well known. In line with previous reports by Eisenhofer and colleagues (4,11-13), in the present study we observed that tumours with mutations in VHL and SDHx mainly produce noradrenaline and those with mutations in RET and NF1 produce both adrenaline and noradrenaline. We also observed that the tumour tissue total catecholamine content was lower in SDHx tumours when compared to RET and sporadic PPGLs. This difference in the catecholamine phenotype has been attributed to the lack of the enzyme phenylethanolamine N-methyltransferase in VHL and SDHx tumours (14). Interestingly, we observed a positive relationship between complex II function, tumour ATP/ADP/AMP content and tumour catecholamine contents suggesting the possibility that differences in energy metabolism might also contribute to the lower tumour tissue catecholamine contents in cluster 1 than in cluster 2 tumours. This inference is supported by the fact that the sequestration of catecholamines into secretory vesicles and re-uptake of catecholamines into chromaffin cells are active energy dependent processes. Sequestration of catecholamines is facilitated by VMATs. The H+ gradient necessary to maintain the activity of VMATs is generated by the ATP-dependent vesicular membrane proton pump (15). Chromaffin granules also contain strikingly high concentrations of ATP due to the activity of vesicular nucleotide transporter (16). This contributes to the stability and enables chromaffin granules to maintain catecholamine stores (17,18). Furthermore, NET responsible for the sodium and chloride (Na+/Cl-) dependent reuptake of extracellular norepinephrine and dopamine is also indirectly dependent on cellular store of ATP. NET functions by coupling the transport of norepinephrine and dopamine with the influx of sodium and chloride (Na+/Cl-). The ion gradients of Na+ and Cl- generated by the Na+/K+-ATPase make this reuptake energetically favorable (17-19). Clearly therefore, energy metabolism has an important role in maintaining the stability of chromaffin granules and thus catecholamine storage. Thereby, it seems possible that genotype-specific differences in the energy metabolism, along with associated differences in expression of various genes encoding components of secretory pathway and exocytotic machinery (20), could in conjunction contribute to genotype-specific differences in tumour catecholamine phenotypic features. In in vitro experiments with cells silenced for SDH, it was observed that the accumulation of succinate leads to stabilization of HIF-1α (21). In the present study, for the first time we provide ex vivo evidence for accumulation of succinate in SDHx tumours. This was later confirmed in a large cohort of patients by Richter and colleagues (22). They further extended the study to detect various Krebs cycle intermediates in PPGLs and studied the use of succinate: fumarate ratio as a

115

Discussion

6

diagnostic marker to detect SDHx mutations in PPGL patients. Recently, Favier and colleagues reported in vivo detection of succinate using in vivo NMR spectroscopy further confirming our initial report (23) (24). We further extended the results of genotype-specific differences in metabolite levels using multiple complementary analytical methods to further investigate genotype-specific differences in tumour metabolite profile. The results of this study presented in chapter 2 demonstrate that the differences in metabolite profiles can be used to differentiate the genotypes using unsupervised statistical methods. In this study, we used two different metabolite profiling platforms namely, NMR and LC-tandem-MS. We used NMR as an untargeted metabolite profiling technique allowing detection of the vast majority of proton containing compounds in the low micromolar or higher concentration range. For targeted metabolite profiling, we used LC-tandem MS to test the presence of genotype-spe-cific differences in the concentration of various purines in the tumours. We used PCA as an unsupervised method for interpretation of NMR spectra and to classify the samples in genotype-specific groups on the basis of their metabolite profile. The peak positions contributing to maximum variability in the first three principal components included epinephrine, succinate, lactic acid and high energy phosphates of adenosine. The lactic acid levels however need to be interpreted with caution as the lactate levels in resected tumour samples are not always reflective of the actual metabolism. One of the striking observations of NMR spectra of SDH-related tumours was the lack of peaks of high energy phosphates of adenosine which suggested that there could be a total lack of adenine content in SDHx tumours. This hypothesis was verified by assessing the content of purines in PPGLs using LC-tandem MS. As described in chapter 2, we observed that the levels of adenine and adenosine were lower in SDHx tumours when compared to RET-related tumours. Inosine levels were lower in cluster 1 tumours compared to cluster 2 tumours in contrast to hypoxanthine levels which were lower in cluster 2 tumours when compared to cluster 1 tumours. We hypothesize that in the de novo synthesis of the purine nucleotides, possibly the activity of inosine monophosphate (IMP) dehydrogenase is higher in SDH tumours while in the RET-related ones, adenylosuccinate synthase and/or lyase activities are higher. Thus, the IMP that is formed by the de novo pathway is committed to either of the purine biosynthetic branches in a geno-type-specific way (Figure 3, Chapter 2). In this study, for the first time we report the presence of N-Acetyl aspartic acid (NAA) in the peripheral nervous system. NAA is synthesized in mitochondria through acetylation of aspartate by L-Asparatate N-acetyl transferase, a membrane bound enzyme, in the neuronal tissue (25). Mitochondrial NAA production is dependent on functioning of respiratory chain enzyme complexes

116

and oxidative phosphorylation. Inhibition of these enzymes leads to significant decrease in mitochondrial NAA and ATP production in isolated rat brain mitochondria (26). This relates well with our observation of strong positive correlation between NAA content and respiratory chain complex II activity and NAA and ATP/ADP/AMP contents in PPGLs. Similar to patients with mitochondrial encephalomyelopathies, we observed low NAA/total creatine ratio in tumours with compromised mitochondrial function. Thus, the NAA levels in SDH and VHL tumours are directly linked to the energetic states of the tumour cell mitochondria. The metabolic features that characterize the SDHx and VHL tumours include lack of epinephrine, NAA and ATP/ADP/AMP while low levels of creatine and presence of lactate are more specific to VHL tumours. RET and NF-1 tumours on the other hand are distinguished by high levels of epinephrine, NAA, ATP/ADP/AMP, and creatine.

6.2 Hypoxia and metastatic potential

Stabilization of HIF-1α and 2α in cluster 1 PPGLs is caused by decreased pVHL-dependent proteasomal degradation by the E3 ubiquitin ligase complex or inhibition of prolyl hydroxylase activity due to succinate accumulation caused by mutations in the SDH subunit. This leads to increased transcription of HIF-dependent genes involved in tumourigenesis and angiogenesis (21). On the other hand, sustained increase in HIF-1α leads to decrease in HIF-1α mRNA. In vitro experiments in renal cell carcinoma cells have yielded an insight into oxygen dependent regulation of HIF-1α mRNA levels (27,28). As discussed in chapter 3, similar to these findings we observed that increased aHIF expression in parallel with VEGF in SDH- and VHL-related PGL fits a pattern of a chronic (pseudo)hypoxic state. This suggests that aHIF can serve as a surrogate marker of chronic HIF-1α activation, taking into account that measurement of the HIF-1α protein itself is cumbersome. Besides its involvement in expression of genes which lead to tumourigenesis, HIF-1α is also involved in stabilization of p53. aHIF expression leads to reduced levels of HIF-1α thus reducing the amount of p53 which is a contributing factor leading to aggressive tumour phenotype which has been demonstrated in other cancer types. As described in chapter 3, in PPGLs as well, aHIF expression in primary tumours turned out to be a predictor of decreased me-tastasis-free survival. There are currently no reliable histological criteria for malignant PGL besides the presence of metastatic lesions in locations where chromaffin tissue is normally absent (29). The 5-year survival of patients with metastatic PGL is nearly 50% (30,31), and no cure is available. It is therefore critical to identify biomarkers of malignancy such as aHIF. Besides aHIF, VEGF was also

117

Discussion

6

negatively correlated with metastasis-free survival which highlights the importance of anti-angiogenic therapy in PPGLs.

6.3 Role of hypoxia and altered energy metabolism in 18F-FDG uptake

Activation of the hypoxic-angiogenic pathway through stabilization of HIF-1α and -2α, leads to increased expression of genes involved in glucose metabolism (glucose transporters and hexokinases), angiogenesis, cell survival and cell motility (1,21). High uptake of 18F-FDG by SDHx-related tumours has been suggested to be a reflection of the Warburg effect. Thus, we examined the ex vivo expression of markers of the Warburg effect in PPGLs using immunohistochemical staining and their correlation with uptake of in vivo 18F-FDG on PET/CT scans. As reported in chapter 4, we observed an increased expression of Hexokinase 2 in SDHx-related tumours when compared to sporadic and cluster 2 tumours. 18F-FDG uptake of any tumour is dependent on amount of 18F-FDG that enters the cell through facilitative glucose transport via GLUTs as well as its phosphorylation through hexokinase which prevents escape of 18F-FDG from the tumour cell. Tumoural blood flow which brings in the nutrients and oxygen also plays a critical role in uptake of 18F-FDG. However, in pseudohypoxia, activation of the hypoxic angiogenic pathway through stabilization of HIF-1α and -2α leads to uncoupling of the blood flow and metabolism implying hypoxic stimulation of glucose metabolism in the presence of adequate oxygen supply. We observed no increase in the expression of GLUT-1 or -3 transporters in cluster 1 tumours. However, an increased expression of hexokinase 2 and 3 and VEGF which correlated with the 18F-FDG uptake values was observed. Also, higher mean vascular perimeter in SDH-related tumours as determined by CD-34 staining could contribute to increased 18F-FDG uptake supporting the idea that permeability and functionality of the vessels is more critical for the uptake than microvessel density.

6.4 Conclusion

This thesis aimed at improving the understanding of the impact of genotype-spe-cific differences in energy metabolism of PPGLs on catecholamine metabolism, imaging and metastatic potential. The studies presented in chapter 1 and 2 concluded that based on differences in metabolite profiles, the PPGLs can be classified into different genotypes in an unsupervised fashion. Alterations in energy metabolism in SDH- and VHL-related tumours can have far reaching

118

impact on catecholamine content and secretion, purine metabolism and amino acid metabolism. Activation of hypoxic angiogenic pathway which leads to subversion of energy metabolism can also be used to predict metastatic potential. As discussed in chapter 3, aHIF and VEGF expressed as a result of activation of hypoxic-angiogenic pathway can be used as markers of metastatic potential. Subversion of energy metabolism by activation of hypoxic-angiogenic pathway is reflected in increased 18F-FDG uptake and increased 18F-FDG uptake is a result of increased expression of hexokinase 2 and 3 and probably increased vascular permeability and higher mean vascular perimeter.

6.5 Future Perspectives

Study of energy metabolism is a relatively new area of research in the study of pathophysiology of PPGL. The studies described in the present thesis as well as recent reports by others have opened new possibilities in understanding patho-physiology of SDH- and VHL-related PPGLs. A detailed study of alterations in energy metabolism using in vitro and in vivo metabolite labeling approaches will help in improved understanding of the pathophysiology. Genotype-specific differences in metabolism can be also used to direct alternative diagnostic and therapeutic options. Detection of succinate using in vivo NMR spectroscopy can be used for localization of SDHx related tumours. Also, detection of succinate, NAA and purines in tumour tissue using 1H NMR spectroscopy or LC-MS can help in inexpensive classification of PPGLs into different genotypes. Inhibitors of glycolysis which target hexokinases like 2-deoxy glucose, and 3-bromo pyruvate can be of interest especially using targeted drug delivery in SDH-related tumours as they show an increased expression of hexokinase and glycolytic activity as measured by increased 18F-FDG avidity (32,33). Shunting of inosine monophosphate to guanosine monophosphate synthesis could also be an attractive therapeutic target as it appears to be the active purine biosynthetic route in SDH-related tumours. Thus, studies on metabolism of PPGLs have opened up new diagnostic and therapeutic options especially in combination with targeted drug delivery.

119

Discussion

6

6.6 References1. Dahia PL. Transcription association of VHL and SDH mutations link hypoxia and oxidoreductase

signals in pheochromocytomas. Ann N Y Acad Sci. 2006;1073:208-220.2. Favier J, Briere JJ, Burnichon N, Riviere J, Vescovo L, Benit P, Giscos-Douriez I, De Reynies A, Bertherat

J, Badoual C, Tissier F, Amar L, Libe R, Plouin PF, Jeunemaitre X, Rustin P, Gimenez-Roqueplo AP. The Warburg effect is genetically determined in inherited pheochromocytomas. PloS one. 2009;4(9):e7094.

3. Rapizzi E, Ercolino T, Canu L, Giache V, Francalanci M, Pratesi C, Valeri A, Mannelli M. Mitochondrial function and content in pheochromocytoma/paraganglioma of succinate dehydrogenase mutation carriers. Endocr Relat Cancer. 2012;19(3):261-269.

4. Eisenhofer G, Pacak K, Huynh TT, Qin N, Bratslavsky G, Linehan WM, Mannelli M, Friberg P, Grebe SK, Timmers HJ, Bornstein SR, Lenders JW. Catecholamine metabolomic and secretory phenotypes in phaeochromocytoma. Endocr Relat Cancer. 2011;18(1):97-111.

5. Timmers HJ, Pacak K, Huynh TT, Abu-Asab M, Tsokos M, Merino MJ, Baysal BE, Adams KT, Eisenhofer G. Biochemically silent abdominal paragangliomas in patients with mutations in the succinate dehydrogenase subunit B gene. The Journal of clinical endocrinology and metabolism. 2008;93(12):4826-4832.

6. Timmers HJ, Chen CC, Carrasquillo JA, Whatley M, Ling A, Havekes B, Eisenhofer G, Martiniova L, Adams KT, Pacak K. Comparison of 18F-fluoro-L-DOPA, 18F-fluoro-deoxyglucose, and 18F-fluorodopa-mine PET and 123I-MIBG scintigraphy in the localization of pheochromocytoma and paraganglioma. The Journal of clinical endocrinology and metabolism. 2009;94(12):4757-4767.

7. Timmers HJ, Kozupa A, Chen CC, Carrasquillo JA, Ling A, Eisenhofer G, Adams KT, Solis D, Lenders JW, Pacak K. Superiority of fluorodeoxyglucose positron emission tomography to other functional imaging techniques in the evaluation of metastatic SDHB-associated pheochromocytoma and paraganglioma. J Clin Oncol. 2007;25(16):2262-2269.

8. Gimenez-Roqueplo AP, Favier J, Rustin P, Mourad JJ, Plouin PF, Corvol P, Rotig A, Jeunemaitre X. The R22X mutation of the SDHD gene in hereditary paraganglioma abolishes the enzymatic activity of complex II in the mitochondrial respiratory chain and activates the hypoxia pathway. Am J Hum Genet. 2001;69(6):1186-1197.

9. Gimenez-Roqueplo AP, Favier J, Rustin P, Rieubland C, Kerlan V, Plouin PF, Rotig A, Jeunemaitre X. Functional consequences of a SDHB gene mutation in an apparently sporadic pheochromocytoma. The Journal of clinical endocrinology and metabolism. 2002;87(10):4771-4774.

10. Fliedner SM, Kaludercic N, Jiang XS, Hansikova H, Hajkova Z, Sladkova J, Limpuangthip A, Backlund PS, Wesley R, Martiniova L, Jochmanova I, Lendvai NK, Breza J, Yergey AL, Paolocci N, Tischler AS, Zeman J, Porter FD, Lehnert H, Pacak K. Warburg effect’s manifestation in aggressive pheochromocy-tomas and paragangliomas: insights from a mouse cell model applied to human tumour tissue. PloS one. 2012;7(7):e40949.

11. Eisenhofer G, Lenders JW, Goldstein DS, Mannelli M, Csako G, Walther MM, Brouwers FM, Pacak K. Pheochromocytoma catecholamine phenotypes and prediction of tumour size and location by use of plasma free metanephrines. Clin Chem. 2005;51(4):735-744.

12. Eisenhofer G, Lenders JW, Siegert G, Bornstein SR, Friberg P, Milosevic D, Mannelli M, Linehan WM, Adams K, Timmers HJ, Pacak K. Plasma methoxytyramine: a novel biomarker of metastatic pheo-chromocytoma and paraganglioma in relation to established risk factors of tumour size, location and SDHB mutation status. Eur J Cancer.48(11):1739-1749.

13. Eisenhofer G, Lenders JW, Timmers H, Mannelli M, Grebe SK, Hofbauer LC, Bornstein SR, Tiebel O, Adams K, Bratslavsky G, Linehan WM, Pacak K. Measurements of plasma methoxytyramine, norme-tanephrine, and metanephrine as discriminators of different hereditary forms of pheochromocyto-ma. Clin Chem. 2011;57(3):411-420.

14. Eisenhofer G, Walther MM, Huynh TT, Li ST, Bornstein SR, Vortmeyer A, Mannelli M, Goldstein DS, Linehan WM, Lenders JW, Pacak K. Pheochromocytomas in von Hippel-Lindau syndrome and multiple endocrine neoplasia type 2 display distinct biochemical and clinical phenotypes. The Journal of clinical endocrinology and metabolism. 2001;86(5):1999-2008.

120

15. Henry JP, Sagne C, Bedet C, Gasnier B. The vesicular monoamine transporter: from chromaffin granule to brain. Neurochem Int. 1998;32(3):227-246.

16. Sawada K, Echigo N, Juge N, Miyaji T, Otsuka M, Omote H, Yamamoto A, Moriyama Y. Identification of a vesicular nucleotide transporter. Proc Natl Acad Sci U S A. 2008;105(15):5683-5686.

17. Galli A, DeFelice LJ, Duke BJ, Moore KR, Blakely RD. Sodium-dependent norepinephrine-induced currents in norepinephrine-transporter-transfected HEK-293 cells blocked by cocaine and antide-pressants. J Exp Biol. 1995;198(Pt 10):2197-2212.

18. Winkler H. The composition of adrenal chromaffin granules: an assessment of controversial results. Neuroscience. 1976;1(2):65-80.

19. Torres GE, Gainetdinov RR, Caron MG. Plasma membrane monoamine transporters: structure, regulation and function. Nat Rev Neurosci. 2003;4(1):13-25.

20. Eisenhofer G, Huynh TT, Elkahloun A, Morris JC, Bratslavsky G, Linehan WM, Zhuang Z, Balgley BM, Lee CS, Mannelli M, Lenders JW, Bornstein SR, Pacak K. Differential expression of the regulated catecholamine secretory pathway in different hereditary forms of pheochromocytoma. Am J Physiol Endocrinol Metab. 2008;295(5):E1223-1233.

21. Selak MA, Armour SM, MacKenzie ED, Boulahbel H, Watson DG, Mansfield KD, Pan Y, Simon MC, Thompson CB, Gottlieb E. Succinate links TCA cycle dysfunction to oncogenesis by inhibiting HIF-alpha prolyl hydroxylase. Cancer Cell. 2005;7(1):77-85.

22. Richter S, Peitzsch M, Rapizzi E, Lenders JW, Qin N, de Cubas AA, Schiavi F, Rao JU, Beuschlein F, Quinkler M, Timmers HJ, Opocher G, Mannelli M, Pacak K, Robledo M, Eisenhofer G. Krebs cycle metabolite profiling for identification and stratification of pheochromocytomas/paragangliomas due to succinate dehydrogenase deficiency. The Journal of clinical endocrinology and metabolism. 2014;99(10):3903-3911.

23. Lussey-Lepoutre C, Bellucci A, Morin A, Buffet A, Amar L, Janin M, Ottolenghi C, Zinzindohoue F, Autret G, Burnichon N, Robidel E, Banting B, Fontaine S, Cuenod CA, Benit P, Rustin P, Halimi P, Fournier L, Gimenez-Roqueplo AP, Favier J, Tavitian B. In vivo detection of succinate by magnetic resonance spectroscopy as a hallmark of SDHx mutations in paraganglioma. Clin Cancer Res. 2015.

24. Rao JU, Engelke UF, Rodenburg RJ, Wevers RA, Pacak K, Eisenhofer G, Qin N, Kusters B, Goudswaard AG, Lenders JW, Hermus AR, Mensenkamp AR, Kunst HP, Sweep FC, Timmers HJ. Genotype-specific abnormalities in mitochondrial function associate with distinct profiles of energy metabolism and catecholamine content in pheochromocytoma and paraganglioma. Clin Cancer Res. 2013;19(14):3787-3795.

25. Patel TB, Clark JB. Synthesis of N-acetyl-L-aspartate by rat brain mitochondria and its involvement in mitochondrial/cytosolic carbon transport. Biochem J. 1979;184(3):539-546.

26. Bates TE, Strangward M, Keelan J, Davey GP, Munro PM, Clark JB. Inhibition of N-acetylaspartate production: implications for 1H MRS studies in vivo. Neuroreport. 1996;7(8):1397-1400.

27. Thrash-Bingham CA, Tartof KD. aHIF: a natural antisense transcript overexpressed in human renal cancer and during hypoxia. Journal of the National Cancer Institute. 1999;91(2):143-151.

28. Uchida T, Rossignol F, Matthay MA, Mounier R, Couette S, Clottes E, Clerici C. Prolonged hypoxia differentially regulates hypoxia-inducible factor (HIF)-1alpha and HIF-2alpha expression in lung epithelial cells: implication of natural antisense HIF-1alpha. The Journal of biological chemistry. 2004;279(15):14871-14878.

29. Linnoila RI, Keiser HR, Steinberg SM, Lack EE. Histopathology of benign versus malignant sympatho-adrenal paragangliomas: clinicopathologic study of 120 cases including unusual histologic features. Human pathology. 1990;21(11):1168-1180.

30. Eisenhofer G, Bornstein SR, Brouwers FM, Cheung NK, Dahia PL, de Krijger RR, Giordano TJ, Greene LA, Goldstein DS, Lehnert H, Manger WM, Maris JM, Neumann HP, Pacak K, Shulkin BL, Smith DI, Tischler AS, Young WF, Jr. Malignant pheochromocytoma: current status and initiatives for future progress. Endocr Relat Cancer. 2004;11(3):423-436.

31. Timmers HJ, Brouwers FM, Hermus AR, Sweep FC, Verhofstad AA, Verbeek AL, Pacak K, Lenders JW. Metastases but not cardiovascular mortality reduces life expectancy following surgical resection of apparently benign pheochromocytoma. Endocr Relat Cancer. 2008;15(4):1127-1133.

121

Discussion

6

32. Ganapathy-Kanniappan S, Geschwind JF. Tumour glycolysis as a target for cancer therapy: progress and prospects. Molecular cancer. 2013;12:152.

33. Zhao Y, Butler EB, Tan M. Targeting cellular metabolism to improve cancer therapeutics. Cell death & disease. 2013;4:e532.

Summary

7

125

Summary

7

Chapter 1 provides an overview of various aspects of pheochromocytomas and paragangliomas (PPGLs). Pathophysiology and alterations in metabolism in relation to genetic background as well as impact of the above on functional imaging of PPGL are discussed in order to introduce the studies that the thesis comprises. The aim of this thesis was to study the impact of genotype-specific alterations in tumour energy metabolism in PPGLs on diagnostic characteristics and metastatic potential. For this purpose, targeted and untargeted ex-vivo metabolomic investigations of PPGL tumour tissues were compared with clinical characteristics and in vivo functional imaging of patients. In Chapter 2, a study that investigates the relationships between genotype- specific differences in mitochondrial function and catecholamine content in PPGLs is described. Respiratory chain enzyme assays and untargeted metabolite profiling with 1H NMR spectroscopy at 500 MHz were conducted on homogenates of 35 sporadic PPGLs and 59 PPGLs from patients with hereditary mutations in SDHB, SDHD, SDHAF-2, VHL, RET, NF1, and MAX. In SDH-related PPGLs, a significant decrease in complex II activity and a significant increase in complex I, III, and IV enzyme activities were observed when compared to sporadic, RET, and NF1 tumours. Also, a significant increase in citrate synthase enzyme activity (P < 0.0001) was observed in SDH-related PPGLs when compared to sporadic-, VHL-, RET-, and NF1-related tumours. A significant increase in succinate accumulation and a significant decrease in ATP/ADP/AMP levels were observed when compared to sporadic PPGLs and PPGLs of other genotypes. Positive correlations were observed between respiratory chain complex II activity and total catecholamine content and ATP/ADP/AMP and total catecholamine contents in tumour tissues. This study for the first time establishes a relationship between determinants of energy metabolism (activity of respiratory chain enzyme complex II, ATP/ADP/AMP content) and catecholamine content in PPGL tumours. Also, this study for the first time successfully uses NMR spectroscopy to detect catecholamines in PGL tumours and provides ex-vivo evidence for the accumulation of succinate in PPGLs with an SDHx mutation. Chapter 3 is an extension of the previous chapter which investigates in detail the effect of genetic alterations on metabolic networks in PPGLs. In this study, homogenates of 32 sporadic PPGLs and 48 PPGLs from patients with mutations in SDHB, SDHD, SDHAF-2, VHL, RET, and NF-1 were subjected to proton 1H NMR spectroscopy at 500 MHz for untargeted and LC tandem mass spectrometry for targeted metabolite profiling. 1H NMR spectroscopy identified 28 metabolites in PPGLs of which 12 showed genotype-specific differences. Extending the results described in chapter 2, significantly low levels of NAA in SDHx tumours and creatine in VHL tumours were observed compared with sporadics and other genotypes. Positive correlation was observed between NAA and ATP/ADP/AMP

126

content and NAA and complex II activity of PGLs. Targeted purine analysis in PPGLs showed significantly low adenine levels in cluster 1 compared to cluster 2 tumours whereas significantly low levels of guanosine and hypoxanthine were observed in RET tumours compared with SDH tumours. Principal component analysis (PCA) of metabolites could distinguish PPGLs of different genotypes. This study gives a comprehensive picture of alterations in energy metabolism in SDH- and VHL-related PPGLs and establishes the interrelationship of energy metabolism and amino acid and purine metabolism in PPGLs. The study described in chapter 4 tests the hypothesis that genotype-specific overexpression of antisense HIF (aHIF) is associated with an increased metastatic potential. Tumour samples were collected from 87 patients with PPGL. Quantitative PCR was performed for aHIF, vascular endothelial growth factor (VEGF), aquaporin 3, cytochrome b561, p57Kip2, slit homolog 3, and SDHC. Expression was related to mutation status, benign versus malignant tumours, and metastasis-free survival. We found that both aHIF and VEGF were overexpressed in cluster 1 PPGLs and in metastatic tumours. In contrast, slit homolog 3, p57Kip2, cytochrome b561, and SDHC showed overexpression in non-metastatic tumours, whereas no such difference was observed for aquaporin 3. Patients with higher expression levels of aHIF and VEGF had a significantly decreased metastasis-free survival. Higher expression levels of SDHC are correlated with an increased metastasis-free survival. In conclusion, we not only demonstrate a higher expression of VEGF in cluster 1 PPGL, fitting a profile of pseudohypoxia and angiogenesis, but also of aHIF. Moreover, overexpression of aHIF and VEGF marks a higher metastatic potential in PPGL. The study described in chapter 5, tests the hypothesis that increased uptake of 18F-FDG in SDH-related PPGLs is reflective of increased glycolytic activity and is correlated with expression of different proteins involved in glucose uptake and metabolism through the glycolytic pathway. Twenty-seven PPGLs collected from patients with hereditary mutations in SDHB (n = 2), SDHD (n = 3), RET (n = 5), neuro-fibromatosis 1 (n = 1) and MAX (n = 1) and sporadic patients (n = 15) were investigated. Preoperative 18F-FDG PET/CT studies were analyzed, mean and maximum standardized uptake values (SUVs) in manually drawn regions of interest were calculated. The expression of proteins involved in glucose uptake (GLUT-1 and -3, respectively), phosphorylation (HK-1, -2, and -3, respectively), glycolysis (MCT-4), and angiogenesis (VEGF, CD34) were examined in paraffin-embedded tumour tissues using immunohistochemical staining with peroxidase-catalyzed polymerization of diaminobenzidine as a read-out. The expression was correlated with corresponding SUVs. Both maximum and mean SUVs for SDH-related tumours were significantly higher than those for sporadic and other hereditary tumours. The expression of HK-2 and HK-3 was significantly higher in SDH-related

127

Summary

7

PPGLs than in sporadic PPGLs (P = 0.022 and 0.025, respectively). The expression of HK-2 and VEGF was significantly higher in SDH-related PPGLs than in other hereditary PPGLs (P = 0.039 and 0.008, respectively). No statistical differences in the expression were observed for GLUT-1, GLUT-3, and MCT-4. The percentage anti-CD 34 staining and mean vessel perimeters were significantly higher in SDH-related PPGLs than in sporadic tumours. Mean SUVs correlated with the expression of HK-2, HK-3, VEGF, and MCT-4. The activation of aerobic glycolysis in SDH-related PPGLs is associated with increased 18F-FDG accumulation due to accelerated glucose phosphorylation by hexokinases rather than increased expression of glucose transporters. Chapter 6 presents a general discussion of the results and observations presented through the thesis. In addition, future perspectives on the utilization of the results of the studies presented in the thesis to improve the diagnostic and therapeutic options available for PPGLs are discussed.

SamenvattingList of publicationsAcknowledgementsCurriculum Vitae

8

131

Samenvatting

8

Samenvatting

Hoofdstuk 1 betreft een overzicht van diverse aspecten van feochromocytoom en paraganglioom (PPGL). De pathofysiologie en de kenmerken van het tumourcel-metabolisme bij specifieke genetische mutaties worden besproken. Dit in relatie tot de klinische fenotypen en kenmerken bij diagnostisch onderzoek. Het doel van dit promotieonderzoek was het bepalen van de effecten van genotype-specifieke veranderingen in tumour metabolisme op de diagnostische karakteristieken van PPGL en op maligne kenmerken. Voor dit doel werden de metabole profielen van ex-vivo PPGL tumourweefsels onderzocht en vergeleken met klinische karakteris-tieken van patiënten en met in vivo functional imaging resulaten.

In hoofdstuk 2 wordt een studie beschreven naar de relatie tussen genotype-spec-ifieke verschillen in mitochondriële functie en de catecholamine concentraties in PPGL. Enzymatische assays van de ademhalingsketen en metabole studies middels 1H NMR spectroscopie bij 500 MH werden toegepast op homogenaten van 35 sporadische PPGLs en 59 PPGLs van patiënten met erfelijke mutaties in SDHB, SDHD, SDHAF-2, VHL, RET, NF1 en MAX. In SDH-gerelateerde PPGLs werd een forse afname in complex II enzymatische activiteit waargenomen, terwijl de activiteit van complex I, III en IV hoger was ten opzichte van sporadische, VHL-, RET- en NF1-gerelateerde tumouren. Tevens was er een significante toename in de activiteit van het mitochondiële enzym citraat synthase in SDH-gerelateerde PPGLs ten opzicht van VHL-, RET- en NF1-gerelateerde tumouren. Dit ging gepaard met stapeling van de metaboliet succinaat en afname van ATP/ADP/AMP en catecholamineconcentraties in SDH-tumouren. Er werd een positieve correlatie gevonden tussen complex II activiteit enerzijds en ATP/ADP/AMP en catecholamine concentraties anderzijds. Dit is de eerste studie waarbij NMR spectroscopie met succes werd toegepast voor het meten van catecholamines in PPGL en waarbij het ex-vivo bewijs geleverd wordt voor de stapeling van succinaat in SDH-deficiënte tumouren.

Het onderzoek beschreven in hoofdstuk 3 is een vervolg op de studie in hoofdstuk 2, waarbij de effecten van genetische mutaties op metabole netwerken in PPGL in detail worden bestudeerd. Hiervoor werden de homogenaten van 32 sporadische PPGLs en 48 PPGLs van patiënten met mutaties in SDHB, SDHD, SDHAF-2, VHL, RET en NF-1 onderzocht middels 1H NMR spectroscopie bij 500 MHz, ditmaal ten behoeve van ‘untargeted’ metabole profilering. Aanvullend werd gebruik gemaakt van LC tandem mass spectrometrie voor ‘targeted’ metabole profilering. Met 1H NMR spectroscopie werden 28 metabolieten geïdentificeerd in PPGL waarvan er 12 genotype-specifieke verschillen lieten zien. Hierbij werden onder andere lage

132

NAA concentraties gevonden in SDHx-tumouren en lage creatine concentraties in VHL tumouren in vergelijking met sporadische tumouren en PPGLs met een ander genotype.Er werd een positieve correlatie gevonden tussen NAA en ATP/ADP/AMP concentraties en tussen NAA en complex II activiteit. Targeted analyse van purine metabolieten toonde een lagere adenine concentratie in cluster 1 ten opzichte van cluster 2 tumouren, terwijl juist lagere concentraties van guanosine en hypoxanthine werden gevonden bij RET tumouren ten opzichte van SDH tumouren. Het was mogelijk om met behulp van ‘principle component analysis’ van de metabole profielen PPGLs met verschillend genotypen van elkaar te onderscheiden.Deze studie geeft een compleet overzicht van de veranderingen in energiemetab-olisme bij SDH en VHL-gerelateerde PPGLs en geeft inzicht in de onderlinge relatie tussen energiemetabolisme, aminozuren en purine metabolieten in PPGL.

Bij het onderzoek in hoofdstuk 4 wordt de hypothese getest dat genotype- specifieke overexpressie van antisense HIF (aHIF) geassocieerd is met maligne kenmerken. Tumouren werden verzameld van 87 patiënten met PPGL. Quantitatieve PCR werd uitgevoerd voor aHIF, vascular endothelial growth factor (VEGF), aquaporin 3, cytochrome b561, p57Kip2, slit homolog 3 en SDHC. De expressie werd bestudeerd in relatie tot mutatiestatus, gemetastaseerde versus niet- gemetastaseerde tumouren en metastase-vrije overleving. We vonden dat zowel aHIF als VEGF een hogere expressie toont bij de cluster 1 tumouren en de maligne tumouren. Daarentegen was de expressie van slit homolog 3, p57Kip2, cytochrome b561, en SDHC juist hoger in niet-gemetastasseerde tumouren. Patiënten met tumouren met een hogere expressie van aHIF en VEGF hadden een verminderde metastase-vrije overleving. SDHC expressie correleerde juist met een toegenomen metastase-vrije overleving. Concluderend toont deze studie aan dat naast VEGF ook aHIF sterker tot expressie komt in cluster 1 PPG, hetgeen past in het bekende profiel van pseudohypoxie en angiogenese. Ook zijn aHIF en VEGF geassocieerd met een hogere kans op metastasering van PPGL.

Het onderzoek dat wordt beschreven in hoofdstuk 5 betreft de hypothese dat toegenomen opname van 18F-FDG door SDH-gerelateerde PPGLs een teken is van toegenomen glycolytische activiteit en correleert met de expressie van eiwitten betrokken bij glucose opname en glycolyse. Bij patiënten met PPGL werden 27 tumouren verzameld met de volgende genotypes: SDHB (n = 2), SDHD (n = 3), RET (n = 5), NF 1 en MAX (n = 1). 15 tumouren waren sporadisch. Pre-operatief werden 18F-FDG PET/CT scans verricht waarbij gemiddelde en maximum standard uptake values (SUVs) werden bepaald in de tumouren. De expressie in tumouren van eiwitten betrokken bij glucose uptake (GLUT-1 en -3), phosphorylering (HK-1,

133

Samenvatting

8

-2, and -3), glycolyse (MCT-4) en angiogenese (VEGF, CD34) werd bepaald middels immuunhistochemische kleuring van paraffine-weefselcoupes. Deze expressie werd vergeleken met de corresponderende SUVs. Zoals verwacht waren de SUVs bij SDH-gerelateerde tumouren waren significant hoger dan bij sporadische tumouren en tumouren met een ander genotype. De expressie van HK-2 en HK-3 was significant hoger in SDH-gerelateerde PPGLs dan in sporadische tumouren. De expressie van HK-2 en VEGF was significant hoger in SDH-gerelateerde PPGLs dan in PPGLs met andere genotypes. Er waren echter geen verschillen in de expressie van GLUT-1, GLUT-3 en MCT-4. Het percentage anti-CD 34 positieve cellen en de gemiddelde vaatperimeter waren hoger bij SDH-gerelateerde PPGLs dan bij sporadische tumouren. De gemiddelde SUVs correleerden met de expressie van HK-2, HK-3, VEGF en MCT-4. Daarom leidt activatie van aerobe glycolyse in SDH- gerelateerde PPGLs blijkbaar tot hogere 18F-FDG stapeling via een toename in phosphorylering in plaats van een toename in de expressie van glucosetransporters.

Hoofdstuk 6 bevat de discussie, waarbij de resultaten van het onderzoek in een breder perspectief worden geplaatst. Ook wordt de relevantie van de resultaten voor de (toekomstige) diagnostiek en behandeling van PPGL toegelicht.

135

List of publications

8

List of publications

Standardization and validation of an induced ovulation model system in buffalo cows: Characterization of gene expression changes in the periovulatory follicle Anim Reprod Sci 2009 113(1-4):71-81 Rao JU and Medhamurthy R

Gene Expression Profiling of Preovulatory Follicle in the Buffalo Cow: Effects of Increased IGF-I Concentration on Periovulatory Events PLoS One 2011 6(6): e20754 Rao JU, Shah KB, Puttaiah J, and Medhamurthy R

Overexpression of the natural antisense hypoxia-inducible factor-1alpha transcript is associated with malignant pheochromocytoma/paraganglioma Endocr Relat Cancer 2011 18 (3):323-31 Span PN, Rao JU, Oude Ophuis SB, Lenders JW, Sweep FC, Wesseling P, Kusters B, van Nederveen FH, de Krijger RR, Hermus AR, Timmers HJ.

Staging and functional characterization of pheochromocytoma and paraganglioma by 18Ffluorodeoxyglucose (18F-FDG) positron emission tomography J Natl Cancer Inst 2012 104 (9):700-8 Timmers HJ, Chen CC, Carrasquillo JA, Whatley M, Ling A, Eisenhofer G, King KS, Rao JU, Wesley RA, Adams KT, Pacak K.

Genotype-specific abnormalities in mitochondrial function associate with distinct profiles of energy metabolism and catecholamine content in pheochromocytoma and paraganglioma Clin Cancer Res 2013 19(14):3787-95 Rao JU, Engelke UF, Rodenburg RJ, Wevers RA, Pacak K, Eisenhofer G, Qin N, Kusters B, Goudswaard AG, Lenders JW, Hermus AR, Mensenkamp AR, Kunst HPM, Sweep FC, Timmers HJ.

Correlation of markers of Warburg effect with the uptake of 18F-fluorodeoxy-glucose (18FFDG) in pheochromocytoma and paraganglioma J Nucl Med 2014 55(8):1253-1259 van Berkel A*, Rao JU*, Kusters B, Demir T, Visser E, Mensenkamp AR, van der Laak J, Sweep FC, Wevers RA, Hermus AR, Langenhuijsen HJF, Kunst HPM, Pacak K, Gotthardt M, Timmers HJ * Authors contributed equally

136

Krebs cycle metabolite profiling for identification and stratification of pheochromocytomas/paragangliomas due to succinate dehydrogenase deficiency J Clin Endocrinol Metab 2014 99(10):3903-11 Richter S, Peitzsch M, Rapizzi E, Lenders JW, Qin N, de Cubas AA, Schiavi F, Rao JU, Beuschlein F, Quinkler M, Timmers HJ, Opocher G, Mannelli M, Pacak K, Robledo M, Eisenhofer G.

Genotype-specific differences in metabolic profiles of pheochromocytoma and paraganglioma as assessed by 1H-NMR spectroscopy and HPLC tandem mass spectrometry J Clin Endocrinol Metab 2015 100(2):E214-22 Rao JU, Engelke UF, Sweep FC, Pacak K, Kusters B, Goudswaard AG, Hermus AR, Mensenkamp AR, Eisenhofer G, Qin N, Richter S, Kunst HPM, Timmers HJ, Wevers RA

Semi-quantitative 123I-metaiodobenzylguanidine scintigraphy to distinguish pheochromocytoma and paraganglioma from physiological adrenal uptake and its correlation with genotype-dependent expression of catecholamine transporters J Nucl Med 2015 56(6):839-46 van Berkel A, Rao JU, Lenders J, Pellegata N, Kusters B, Piscaer I, Hermus A, Plantinga T, Langenhuijsen H, Vriens D, Janssen M, Gotthardt M, Timmers H.

Imaging the intratumoural-peritumoural extracellular pH gradient of gliomas NMR in Biomed 2016 29(3):309-19. Coman D, Huang Y, Rao JU, De Feyter HM, Rothman DL, Juchem C, Hyder F

A ketogenic diet increases transport and oxidation of ketone bodies in RG2 and 9L gliomas without affecting tumour growth. Neuro Oncol 2016 18(8):1079-87. De Feyter HM, Behar KL, Rao JU, Madden- Hennessey K, Ip KL, Hyder F, Drewes LR, Geschwind JF, de Graaf RA, Rothman DL

Hypofrontality and posterior hyperactivity in early schizophrenia: Multi-modal imaging and behavior in a preclinical model Epub Biological Psychiatry 2016 Kaneko G, Sanganahalli BG, Groman SM, Wang H, Coman D, Rao JU, Herman P, Jiang L, Rich K, de Graaf R, Hyder F DOI: 10.1016/j.biopsych.2016.05.019

137

List of publications

8

Towards longitudinal mapping of extracellular pH in gliomas NMR in Biomed 2016 29(10):1354-72. Huang Y, Coman D, Herman P, Rao JU, Hyder F

139

Acknowledgements

8

Acknowledgements

It gives me great pleasure to have completed this thesis and to pause for a moment to take a look back at the journey. The work presented in this thesis is the result of a strong collaborative spirit and help of many, many people who were directly and indirectly involved in bringing the projects to fruition. I would like to take this occasion to say ‘Thank You’ to all of them.

To begin with, I would like to express my sincere gratitude to my co-promoter and mentor Dr. Henri Timmers. Henri, thank you for trusting me, giving me freedom, shaping my thinking, for being a great teacher and exemplary mentor. You have played a crucial role in moulding my scientific career and shaping my ability to work with people in highly interdisciplinary teams. By observing the way you went about planning and executing projects, I learnt a lot and these are the most valuable lessons that will help me throughout my career. Thank you for giving me an opportunity to work on the projects which I thoroughly enjoyed. My love for understanding the biological basis of imaging contrast mechanisms and cancer metabolism stemmed from my work at Radboud and will stay with me. Over the past 3 years when it got really tough for me, you have been very supportive, patient and understanding. I really appreciate it. Thank you for everything.

The work presented in this thesis would not have been possible without the able guidance and mentoring provided by my promoters. Prof. Wevers, thank you for all the guidance, support and advice on the projects. It was my good fortune that I had you as my promotor and you agreed for a bi-weekly meeting and came walking all the way from LGEM on the other end of hospital. I used to look forward to every meeting and each of those meetings have helped in shaping the projects and papers I worked on at Radboud. Without your crucial inputs none of the projects described in the thesis would have been realized. I owe to you my knowledge on tumour metabolism and the critical aspects of designing translational studies.Prof. Sweep, thank you for the guidance and support and providing for me a place in the dept. of Laboratory Medicine. Thank you for the bi-weekly meetings which provided a platform for discussion of all the scientific, technical and logistical issues related to the projects. Your help in defining the projects and providing me with the entire infrastructure to carry them out was crucial for their completion.

Prof. Hermus, thank you for your support and making me feel a part of division of Endocrinology, dept of Internal Medicine.Prof. Lenders, thank you for your support and advice with the project and your

140

critical comments and active participation in the meetings.Dr. Karel Pacak and Prof. Graeme Eisenhofer, thank you for a great collaboration and for pooling in the samples for all the studies. Thank you for the guidance and help in validating the studies. It was a great learning experience working with your research groups.Prof. Gotthardt and Prof. Boerman, thank you for your guidance with analysis of 18F-FDG PET data and comparison of 18F-FDG uptake with expression of glycolytic pathway proteins.Richard, thank you for your collaboration and support with the respiratory chain enzyme assays. Udo, thank you for all the help with NMR spectroscopy and teaching me related data analysis. Thank you, Anne and Angelina for help with the related experiments.Benno, thank you for the assessment of histological sections of PPGLs and teaching me pathology of PPGLs and helping me understand the slides and shaping my thinking on cancer.Jeroen, thank you for introducing me to the world of quantitative histology. It was a great learning experience.Arjen, thank you for the help with genetic testing of patients, especially the additional testing we requested for several patients as new tests became available and for providing us with the reports. Thank you, Egbert for allowing me the use of histology lab at dept. of Urology and Mirjam, Cathelijne and Jannekke your help with antibodies and protocols is greatly appreciated.

Anouk, I was fortunate to find a great colleague and friend like you. Finally, when Henri had another PhD student a few months before I left, I felt so happy to be able to share my enthusiasm for tumour metabolism with another soul. Together, we worked on the 18F-FDG uptake project and I thoroughly enjoyed working with you.Romana, thank you for being a good friend. I will always remember your bright encouraging smile!Thank you, Kim and Annenienke for your friendship and everyone at division of Endocrinology for your help and all the great times. Thank you, Theo and Leo for help with access to Internal Medicine laboratory.Thank you, Doorlene for being a great friend and helping me with the blood sample and tumour tissue biobank maintenance. Nicolai, thank you for being a great friend, for clearing many of my doubts on statistics and for the great conversations ranging from science to politics and to being an expat. Annekke, thank you for helping me with the VEGF antibody related data. Joop, thank you for help with qPCR and guiding me with shipping samples. I would like to thank all the members of the erstwhile, Afdeling Chemische Endocrinologie, it was

141

Acknowledgements

8

great knowing you all and thank you for all the help I got from you and the great times. I have missed you all in the U.S.

During my stay in the Netherlands, at Bunnik and later in Nijmegen, I made lots of friends. I thank them all for the wonderful times I had there. A big thank you to Medha, Raja, Rama, Deepa, Akshata, Sai, Suresh, Shankar and Sonu. My neighbours in Nijmegen, Sahana and Jagadish and their cute son, Ishaan will always remain dear to me. I thank you all for the help I received from you people and the great times we had together!

I would also like to thank my friends at Yale who supported me as I juggled my postdoc project there with the completion of this thesis. Thank you, Basav and Lata for your friendship and the immense help and support you provided. You were always there for us and we miss you people especially Krish who is growing up to be a lovely little boy. Thank you, Maggie and John for being such wonderful friends and helpful colleagues. Thank you, Fahmeed, Daniel and Golden for your understanding and support.

This thesis would not have been possible without the incredible support I received from my family especially after my son Nachiketa whom I call Minchu was born. I would like to thank my in laws who stepped in to help me when I needed it badly. Amma, Anna, I will always be grateful for helping me out when I needed it most.Appa, Amma, without your encouragement, I would not have reached here. You did everything possible to support me and always encouraged me to do my best! I can never thank you enough for my life is your gift.

Throughout these years there has been one person standing as a strong pillar of support, no matter what. Thank you Umesh, for being my beloved friend, philosopher and guide and a great husband. Without your support, I could not have completed this thesis or advanced in my career. Sorry, it did take a long time and pressures of two careers has its challenges. But together, we have learnt to smile at the face of each challenge and gladly welcome it.

My dearest Minchu, you were born when I was in the middle of this project which was going full swing. Many nights you have seen me awake next to your bed. Sorry, the brightness of my computer screen might have disturbed your sleep, you searched for Amma, and she was in lab and when you wanted to play she was too tired. When she was tired you always understood and made her laugh. You never fail to put a smile on my face! I am privileged to be your mother. Thank you.

142

This work, is not mine, it is your grace Sri Guru who, through all my mentors and friends and family members, have guided me. Many a times when I had a paradigm shift while thinking about a project or realized how wonderous a thing this living mechanism is, I am filled with immense joy and I bow down to thee who is the source of knowing from where all the knowledge emerges. To you I dedicate all that you enabled me to learn and do.

143

Acknowledgements

8

145

CV

8

Curriculum Vitae

Jyotsna Upendra Rao was born on December 1st, 1980 in Udupi, a small town in Karnataka, India. She finished her Pre-University education at Malnad Education Society, Chikmagalur in 1998 and studied her undergraduate course (B.Sc.) in Biology and Chemistry at Mahatma Gandhi Memorial College, Udupi and graduated in 2001 with distinction. Following this, she pursued M.Sc. in Biochemistry at University of Mysore, Mysore and obtained her Master’s degree in 2003 with distinction. In 2003, she received Ministry of Human Resources and Development scholarship to pursue her doctoral studies at the Indian Institute of Science, Bangalore, India in Molecular Endocrinology. She defended her PhD thesis entitled ‘Studies on growth and development of the ovarian dominant follicle in mono - vulatory species: Analysis of transcriptional changes and factors influencing periovulatory events’ in 2009 and was awarded a doctorate degree in Biology. In 2010, she joined depts. of Internal Medicine and Laboratory Medicine, Radboud University Medical Center as a PhD student under the supervision of Dr. Henri Timmers, Prof. Ron Wevers, Prof. Fred Sweep and Prof. Ad Hermus where she did the work described in this thesis. After moving to US, in 2014, she joined the dept. of Radiology and Biomedical Imaging, Yale School of Medicine, as a postdoctoral fellow. She also worked as a MIT(USA)-DBT(Govt. of India) alliance Visiting Scholar at Manipal University, Manipal, India from Dec 2015 – Feb 2016. After successfully finishing her postdoctoral project there in July 2016, she joined Cardiff University School of Medicine, Cardiff, UK as a visiting research scientist where she is currently trying to develop her independent research project on use of 18F-FDOPA imaging in glioma patients. She will continue her research on tumor metabolism in Cancer Research UK Cambridge Institute, University of Cambridge starting in June 2017.