G Model

YSCDB 1773 1–8 ARTICLE IN PRESS

Seminars in Cell & Developmental Biology xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

Seminars in Cell & Developmental Biology

journal homepage: www.elsevier.com/locate/semcdb

1 Review

2 Models of parent-of-origin tumorigenesis in hereditary

3 paraganglioma
4 Q1 Attje S. Hoekstra a , Peter Devilee a,b , Jean-Pierre Bayley a,∗
a
5 Department of Human Genetics, Leiden University Medical Center, Leiden, The Netherlands
b
6 Department of Pathology, Leiden University Medical Center, Leiden, The Netherlands
7

8
22 a r t i c l e i n f o a b s t r a c t
9
10 Article history: Paraganglioma and pheochromocytoma are neuroendocrine tumors that originate from either the sym-
11 Received 20 March 2015 pathetic or the parasympathetic branches of the autonomic nervous system. Although 14 different genes
12 Received in revised form 27 May 2015 have been linked to paraganglioma/pheochromocytoma, a subgroup of these genes is associated with
13 Accepted 28 May 2015
hereditary paraganglioma–pheochromocytoma, the genes related to mitochondrial succinate dehydro-
14 Available online xxx
genase (SDH) including SDHA, SDHB, SDHC, SDHD and the assembly factor SDHAF2. Unlike mutations in
15
other SDH subunit genes, mutations in SDHD and SDHAF2 show a remarkable parent-of-origin dependent
16 Keywords:
tumorigenesis in which tumor formation almost exclusively occurs following paternal transmission of
17 Paraganglioma
18 Pheochromocytoma the mutation. To date, three different models have sought to explain the striking inheritance pattern seen
19 SDHAF2 in SDHD and SDHAF2-linked families. Despite the fact that the models suffer to varying degrees from a
20 SDHD lack of experimental verification, all three models have made some attempt to incorporate current data
21 MAX and understanding of this phenomenon. In this review, we discuss our present understanding of this
phenomenon and describe the three models that seek to explain the inheritance pattern in SDHD and
SDHAF2-linked families.
© 2015 Elsevier Ltd. All rights reserved.

23 Contents

24 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
25 2. Associated genes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
26 3. Succinate dehydrogenase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
27 4. MAX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
28 5. Inheritance pattern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
29 6. Models of parent-of-origin tumorigenesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
30 6.1. The Hensen model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
31 6.2. The Muller threshold model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
32 6.3. The Baysal methylated boundary element model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
33 7. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
34 Acknowledgement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
35 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
36
39

37Q3 1. Introduction Pheochromocytomas are closely related neuroendocrine tumors 40

Q4 but are exclusively associated with the sympathetic nervous sys- 41

38 The neuroendocrine tumor, paraganglioma, is associated with tem. Parasympathetic paragangliomas occur most commonly in the 42

both the sympathetic and parasympathetic nervous systems. head and neck region and most frequently arise in the carotid body 43

at the bifurcation of the carotid artery, while pheochromocytomas 44

arise in the adrenal medulla and sympathetic paragangliomas 45

∗ Corresponding author at: Department of Human Genetics, Leiden Univer-

occur elsewhere in the abdomen and thorax [1]. Paragangliomas 46

Q2 sity Medical Centre, PZ S-04, P.O. Box 9600, 2300 RC Leiden, The Netherlands.
of the head and neck usually show relatively mild symptoms and 47

Tel.: +31 715269512; fax: +31 715268285. display a characteristically indolent tumor progression [2]. Promi- 48

E-mail address: j.p.l.bayley@lumc.nl (J.-P. Bayley). nent characteristics of pheochromocytomas and extra-adrenal 49

http://dx.doi.org/10.1016/j.semcdb.2015.05.011
1084-9521/© 2015 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Hoekstra AS, et al. Models of parent-of-origin tumorigenesis in hereditary paraganglioma. Semin Cell
Dev Biol (2015), http://dx.doi.org/10.1016/j.semcdb.2015.05.011
 G Model
YSCDB 1773 1–8 ARTICLE IN PRESS
2 A.S. Hoekstra et al. / Seminars in Cell & Developmental Biology xxx (2015) xxx–xxx

50 paragangliomas are the hypersecretion of catecholamines and essential for the correct flavination of SDHA and function of the SDH 111

51 elevated blood pressure. Sympathetic paragangliomas also show complex. The first mutation identified in SDHAF2, a missense vari- 112

52 a greater propensity for malignant degeneration compared to ant c.232G>A (p.Gly78Arg), was identified in a large Dutch head and 113

53 pheochromocytomas or head and neck paragangliomas [3]. neck paraganglioma kindred and was shown to result in the loss of 114

SDHA flavination and activity of the SDH complex [6]. 115

A follow-up study in 443 paraganglioma and pheochromocy- 116

54 2. Associated genes
toma patients found no further mutations and demonstrated that 117

SDHAF2 mutations make a very modest contribution to the overall 118

55 Paraganglioma–pheochromocytoma is associated with a het-
genetic burden in these syndromes [24]. A notable characteristic of 119
56 erogeneous collection of genes, with germline or somatic mutations
SDHAF2 mutations is the very high penetrance (i.e., the chance that 120
57 identified in 14 genes to date. These genes belong to a wide range
a mutation carrier will develop a tumor). Of the known mutation 121
58 of functional categories that include kinase receptors (RET), reg-
carriers in the main Dutch kindred, all unaffected individuals are 122
59 ulators of signaling (NF1, HRAS), hypoxia-related factors (VHL,
currently under the age of 45. This high level of penetrance is remi- 123
60 HIF2A), enzymes involved in energy metabolism (SDHA, SDHB,
niscent of SDHD, which also shows very high penetrance compared 124
61 SDHC, SDHD, SDHAF2, FH, MDH2), endosomal signaling factors
to the other SDH genes. 125
62 (TMEM127), vesicle transport/apoptosis (KIF1B), and transcription
63 factors (MAX) [4,5].
64 Although germline mutations in most of these genes lead to 4. MAX 126

65 autosomal dominant inheritance patterns of the disease in families,

66 there are several prominent exceptions: SDHD, SDHAF2 and MAX. MAX (chromosome 14q23) is a basic helix-loop-helix transcrip- 127

67 The SDHD and SDHAF2 genes both encode proteins related to the tion factor that forms a complex with the important oncogene MYC. 128

68 mitochondrial succinate dehydrogenase (SDH) complex and tumor MAX is also found in repressor complexes with other transcrip- 129

69 development occurs almost exclusively in carriers of paternal tion factors, which effectively oppose the function of the MYC-MAX 130

70 mutations [6,7]. Carriers of MAX mutations also show ‘paternal- heterodimer [25]. Germline loss-of-function MAX mutations were 131

71 only’ tumor development, but in contrast to the so-called cluster 1 identified by sequencing the entire exome of 3 unrelated patients 132

72 pseudohypoxia-driven tumors including the SDH genes and VHL, with a family history of pheochromocytoma [8]. Intriguingly, these 133

73 MAX mutations are associated with cluster 2, the PGL/PCC subgroup patients exhibited a loss of heterozygosity of the wild-type allele 134

74 thought to be associated with kinase signaling [5,8]. that was shown to be dependent on uniparental disomy (in this 135

case duplication of the paternal chromosome). An additional 5 136

cases were identified in a follow-up study [8] and in a study of 137

75 3. Succinate dehydrogenase
1694 pheochromocytoma and paraganglioma patients negative for 138

mutations in other known genes, Burnichon et al. showed that 139

76 Mutations in succinate dehydrogenase (SDH) complex genes
germline mutations in MAX were responsible for 1.1% of cases [26]. 140
77 (SDHA, chromosome 5p15; SDHB, chromosome 1p36; SDHC,
Although the total number of patients identified is still limited, 141
78 chromosome 1q23; SDHD, chromosome 11q23; and SDHAF2,
these studies showed exclusively paternal transmission of MAX 142
79 chromosome 11q12) are associated with head and neck paragan-
mutations in affected patients, with tumor formation absent in 143
80 gliomas, extra-adrenal paragangliomas, and pheochromocytomas
individuals with maternally inherited mutations. Furthermore, it 144
81 [6,9–12]. SDHD encodes one of the two membrane-anchoring
was shown that MAX is not imprinted [8]. Although at present 145
82 subunits, while SDHAF2 encodes an accessory factor involved in
there is no explanation for the pattern of inheritance, it is worth 146
83 complex assembly. The SDH genes act as tumor suppressors and all
noting that chromosome 14 harbors several imprinted genes. It is 147
84 appear to follow the Knudson model, showing loss of heterozygos-
possible that loss of the maternally expressed gene 3 (MEG3), an 148
85 ity (LOH) in conjunction with a germline mutation. This results in
imprinted non-coding RNA gene silenced on the paternal allele, 149
86 loss of a protein subunit, which in turn destabilizes the SDH com-
plays a role in MAX tumor development. In accordance with the 150
87 plex as a whole and abolishes enzymatic activity [13]. The catalytic
Knudson model, loss of MAX (and perhaps also MEG3) due to uni- 151
88 subunit consists of two soluble proteins, a flavoprotein (encoded by
parental paternal disomy affecting chromosome 14q is the prelude 152
89 SDHA) and an iron–sulfur protein (SDHB), while subunits encoded
to tumor formation [8]. Alternatively, it is also possible that overex- 153
90 by SDHC and SDHD anchor the complex in the inner mitochondrial
pression of a paternally expressed imprinted gene on chromosome 154
91 membrane and bind ubiquinone.
14 such as RTL1 [27] is essential for tumorigenesis, or that chro- 155
92 SDH, also known as succinate–coenzyme Q reductase, has a dual
mosomal duplication in the tumor is required to delete wild-type 156
93 function as an enzyme of the mitochondrial tricarboxylic acid cycle
MAX while maintaining sufficient expression of an unknown hap- 157
94 and as complex II of the electron transport chain. In the former
loinsufficient gene, loss of which would otherwise limit growth of 158
95 role it oxidizes succinate to fumarate, while in the latter it reduces
tumor cells. 159
96 ubiquinone to ubiquinol, contributing to the generation of ATP by
97 oxidative phosphorylation. SDH inactivation induces an accumu-
98 lation of succinate [14–16] or the generation of reactive oxygen 5. Inheritance pattern 160

99 species (ROS) [17–23], leading to inhibition of ␣-ketoglutarate-

100 dependent HIF prolyl hydroxylases and the activation of hypoxia In the late 1980s, Dr. Andel van der Mey and colleagues at 161

101 inducible factor (HIF). Leiden University Medical Center noticed an unusual pattern of 162

102 Despite the close functional relationship of the SDH proteins, tumor inheritance in families with paraganglioma type 1 (PGL1), 163

103 mutations result in striking differences in both tumor location and with an apparently exclusive development of tumors in the chil- 164

104 clinical phenotype. The molecular basis for this clinical divergence dren of male family members [7]. Clinicians and scientists working 165

105 is not yet understood. at Leiden University Medical Center were well-placed to recog- 166

106 SDHAF2 encodes a gene for a novel protein that acts as an acces- nize the paternal inheritance of PGL1 tumor susceptibility due 167

107 sory/assembly factor for the SDH complex, adding a flavin-adenine to the predominance of PGL1-linked cases, and absence of cases 168

108 dinucleotide (FAD) prosthetic group to form a catalytically active linked to other genes, amongst the numerous affected families in 169

109 SDHA flavoprotein. Initially referred to as SDH5, the succinate dehy- the immediate Leiden area. In a second area of the Netherlands, 170

110 drogenase complex assembly factor 2 (SDHAF2) was shown to be the Nijmegen/Den Bosch region, researchers from the Radboud 171

Please cite this article in press as: Hoekstra AS, et al. Models of parent-of-origin tumorigenesis in hereditary paraganglioma. Semin Cell
Dev Biol (2015), http://dx.doi.org/10.1016/j.semcdb.2015.05.011
 G Model
YSCDB 1773 1–8 ARTICLE IN PRESS
A.S. Hoekstra et al. / Seminars in Cell & Developmental Biology xxx (2015) xxx–xxx 3

172 University Medical Center recognized another focus of paragan- heterogeneity and copy neutral LOH could impair the detection of 234

173 glioma patients that showed a similar phenomenon of exclusively genomic alterations in some tumor samples [42]. In addition, other 235

174 paternal transmission of disease susceptibility [28]. It became clear genetic and epigenetic mechanisms may result in functional loss of 236

175 early on that a second and distinct genetic locus was responsible one or more genes in the 11p region. It is interesting to note that 237

176 for disease in these patients [29]. These gene loci were described as chromosome 11p loss is also a prominent feature in VHL-related 238

177 PGL1 and PGL2, and were later shown to harbor mutations in the PGL/PCC [41,43]. 239

178 SDHD (2000) and SDHAF2 (2009) genes, respectively [6,10]. If the loss of chromosome 11 is important for the development 240

179 The failure of maternally transmitted mutations to initiate of all SDH-related tumors, one prediction of the Hensen model 241

180 tumorigenesis initially suggested that a maternally imprinted gene is that mutations in SDHD and SDHAF2 will display higher pene- 242

181 could be the underlying cause of the tumor [7]. However, follow- trance than mutations in SDHA, B, or C. Tumorigenesis in SDHD and 243

182 ing the identification of SDHD it was demonstrated that the gene SDHAF2 mutation carriers requires only a single somatic genetic 244

183 shows biallelic expression in a range of fetal and adult tissues event (chromosome 11 loss), as opposed to the two events required 245

184 [10,30]. Subsequent studies also failed to uncover evidence for gene in SDHA, B, and C mutation carriers (loss of the respective wild- 246

185 methylation in any tissue analyzed to date [30–32]. It could be type allele, together with independent loss of chromosome 11). 247

186 argued that these findings were not unexpected, due to both the This prediction appears to be borne out by the wide difference in 248

187 ubiquitous and essential nature of the SDH enzyme and because lifetime penetrance between mutations in SDHD (∼90%) [44] and 249

188 mutation carriers with a paternally inherited mutation would expe- SDHAF2 (>95%) [45], compared to SDHB (∼30%) [46–48], and SDHA 250

189 rience a profound deficiency of succinate dehydrogenase activity, and SDHC (both with unknown, but probably very low, penetrance). 251

190 a situation known to cause major developmental defects [33]. Particularly striking is the contrast in penetrance between germline 252

191 Unsurprisingly, homozygous knockout of SDHD in mice results in mutations of SDHA, in which carriers appear to be sporadic cases, 253

192 embryonic lethality [34,35]. Even if one assumes that SDHD shows and its assembly factor, SDHAF2, in which germline mutations have 254

193 tissue-specific or temporally restricted monoallelic expression, the a clearly familial presentation and display almost full penetrance 255

194 complete absence of SDH in any tissue during development would in known pedigrees. As the role of SDHAF2 mutations in tumori- 256

195 probably be incompatible with normal development. genesis is thought to be mediated through the loss of SDHA, SDHA 257

mutations could reasonably be expected to show a similar or even 258

higher penetrance. 259

196 6. Models of parent-of-origin tumorigenesis The Hensen Model is also relevant to the rare cases in which 260

maternal inheritance leads to tumor development. Hensen and col- 261

197 6.1. The Hensen model leagues predicted that “when the SDHD mutation is maternally 262

transmitted, at least two events caused by different chromosomal 263

198 Against a background of a theoretical model of direct maternal mechanisms will be required. . . namely loss of the paternal wild- 264

199 imprinting, an unexpected finding in SDHD-linked paragangliomas type SDHD allele by, for example, mitotic recombination, followed 265

200 was the frequent loss of the maternal copy of chromosome 11 by loss of the recombined paternal chromosome containing the 266

201 [36–39]. Although this conforms to the Knudson two-hit model for paternal 11q23 region and the maternal 11p15 region” [37]. This 267

202 tumor suppressor genes in which the remaining wild-type allele is phenomenon has indeed been observed in at least two cases of 268

203 lost early in tumorigenesis, it is counterintuitive if one assumes maternal inheritance. Both cases were patients with pheochromo- 269

204 that imprinting inactivates the maternal allele. The presence of cytoma, caused by either a c.242CT SDHD mutation [49] or by the 270

205 a paternal parent-of-origin effect in SDHAF2-associated paragan- Dutch founder mutation, c.274G>T, in SDHD [50], both inherited via 271

206 glioma families, together with its absence in families caused by the maternal line. Analysis of microsatellite markers revealed only 272

207 other SDH genes, again argues that their location on chromosome partial loss of the paternal chromosome, with loss of a significant 273

208 11 is an important factor in SDHD- and SDHAF2-related tumorigen- proportion of the maternal chromosome 11 including the p arm 274

209 esis. and centromeric q arm. Both cases showed somatic recombination 275

210 Taken together, these observations suggested that a simple of chromosome 11 in the tumor, resulting in loss of the paternal 276

211 model, in which a paternal mutation is accompanied by a mater- wild-type SDHD gene and the maternal 11p15 region. 277

212 nally imprinted and silenced wild-type allele, was unlikely to

213 explain the inheritance pattern of tumor susceptibility. 6.2. The Muller threshold model 278

214 The role of chromosomal location led to the development of

215 an alternative hypothesis, now known as the ‘Hensen model’ An alternative hypothesis explaining SDHD-linked parental 279

216 [37]. While there is no evidence that either SDHD or SDHAF2 are effects has been advanced by Ulrich Muller, one of the researchers 280

217 imprinted, the main cluster of human imprinted genes is located on involved in the identification of SDHC as the gene underlying 281

218 the short arm of chromosome 11 (11p15). This led to the hypothesis paraganglioma, type 3 (PGL3)[51]. This ‘threshold model’ assumes 282

219 that loss of the maternal copy of chromosome 11 in paragan- a “partial inactivation” of the maternally derived copies of either 283

220 glionic cells of individuals carrying a paternally derived mutation in SDHD or SDHAF2 and requires the existence of a mechanism 284

221 SDHD or SDHAF2 not only completely inactivated wild-type activ- of partial imprinting [52]. Given that all studies to date have 285

222 ity at these loci, but also a critical maternally expressed, paternally found no increase in DNA methylation in the promoter region of 286

223 imprinted gene in the 11p15 region (Fig. 1). It is conceivable that SDHD, partial inactivation of SDHD would require an alternative 287

224 the loss of this as yet unidentified locus (or loci) in 11p15 pro- mechanism. 288

225 vides the necessary conditions in which SDHD-null cells can foster The model proposes that partial inactivation of the maternal 289

226 tumor development. Several genes on chromosome 11 are known allele, together with an inactivating mutation on the paternal allele, 290

227 to be exclusively maternally expressed, including CDKN1C, KCNQ1, results in a state in which some residual activity of SDH is main- 291

228 KCNQ1DN, ZNF215, SLC22A18, PHLDA2, OSBPL5, and H19. A well- tained in cells with a paternally derived mutation (Fig. 2). This 292

229 described gene in the chromosome 11p15.5 region is CDKN1C, the residual activity is sufficient to maintain normal function in para- 293

230 expression of which is frequently deregulated in imprinting disor- ganglia cells over long periods (decades). Functionally, the model 294

231 ders such as Beckwith–Wiedemann Syndrome and Silver–Russell goes on to postulate that ongoing exposure to elevated levels of 295

232 Syndrome [40]. Although it has been shown that the loss of chromo- ROS and succinate (both known to be elevated by SDH gene muta- 296

233 some 11p does not occur in all SDH-related PGL/PCC [37,41], tumor tions), together with chronic hypoxia, favor pathogenic processes 297

Please cite this article in press as: Hoekstra AS, et al. Models of parent-of-origin tumorigenesis in hereditary paraganglioma. Semin Cell
Dev Biol (2015), http://dx.doi.org/10.1016/j.semcdb.2015.05.011
 G Model
YSCDB 1773 1–8 ARTICLE IN PRESS
4 A.S. Hoekstra et al. / Seminars in Cell & Developmental Biology xxx (2015) xxx–xxx

Fig. 1. Hensen model explaining parent-of-origin transmission in SDHD-linked PGL/PCC. Maternal (white) and paternal (gray) chromosomes 11 are depicted. (a) Paternal
transmission of an SDHD mutation together with loss of maternal chromosome 11 targets both the wildtype maternal SDHD allele and a second tumor suppressor gene –
tumor formation is initiated. (b) In the case of transmission of a maternal SDHD mutation, followed by loss of paternal chromosome 11, a second tumor suppressor gene is
still present on the maternal chromosome and tumor formation is suppressed. (c) Upon maternal transmission of a mutated SDHD gene, two events are required for tumor
development: mitotic recombination of the paternal wildtype SDHD allele with a section of the maternal chromosome containing a second tumor suppressor gene, followed
by loss of the recombined chromosome.

298 and chromosomal nondisjunction. This results in loss of maternal Other mechanisms such as inhibition of histone demethylases 314

299 chromosome 11 and causes SDH activity to dip below a permissive [23,53,54] or inhibition of EglN3-mediated apoptosis [55] have 315

300 threshold, triggering the adult-onset of PGL tumorigenesis (Fig. 2b). been linked to SDH-related PGL, which can’t be explained by the 316

301 Conversely, a maternally derived mutation does not have a simi- ‘Muller Threshold’ model. 317

302 larly deleterious effect because sufficient SDH activity and normal A virtue of the ‘Muller Threshold’ hypothesis is that it makes 318

303 function is maintained due to the activity of the paternally derived a clear prediction regarding activities of the respective maternal 319

304 wild-type SDHD allele (Fig. 2c). SDH activity is thus at close to nor- and paternal alleles of the SDHD/SDHAF2 genes and as such exper- 320

305 mal levels (Fig. 2a). In contrast to a paternally derived mutation, imental verification appears to be relatively straightforward. If we 321

306 elevated ROS, succinate and chronic hypoxia are avoided and the assume that gene alleles show differing activities dependent on 322

307 cellular milieu disfavors nondisjunction or loss of the wild-type the parent-of-origin, this could be demonstrated by allele-specific 323

308 allele. expression studies in healthy individuals. The proposed effects 324

309 A weakness of this hypothesis is that it explicitly posits that of low SDH activities on the cellular milieu could be established 325

310 hypoxic stimulation via HIF underlies tumor formation. Whereas through the study of in vitro systems or of primary cells from 326

311 HIF activation is a prominent feature of SDH-related paragan- SDHD/SDHAF2 mutation carriers with either maternal or paternally 327

312 gliomas, it remains unclear how, or indeed whether, it is directly inherited mutations. However, to the best of our knowledge no 328

313 involved in triggering or promoting tumor formation in PGL/PCC. supporting data have appeared to date. 329

Please cite this article in press as: Hoekstra AS, et al. Models of parent-of-origin tumorigenesis in hereditary paraganglioma. Semin Cell
Dev Biol (2015), http://dx.doi.org/10.1016/j.semcdb.2015.05.011
 G Model
YSCDB 1773 1–8 ARTICLE IN PRESS
A.S. Hoekstra et al. / Seminars in Cell & Developmental Biology xxx (2015) xxx–xxx 5

Fig. 2. Threshold model of partial imprinting (inactivation) leading to tumor formation in SDHD-linked PGL/PCC. Maternal (white) and paternal (gray) chromosomes 11 are
depicted. (a) In wildtype cells, SDH activity is high and succinate, ROS and HIF1␣ levels are low due to sufficient combined maternal and paternal expression of SDHD –
tumor formation is not initiated. (b) Paternal transmission of an SDHD mutation eliminates the dominant SDHD allele, decreasing SDH activity and increasing succinate, ROS
and HIF1␣ levels. Chronic cellular hypoxia rises above a threshold (dotted line) beyond which non-disjunction and tumor formation can take place. (c) Following maternal
transmission of an SDHD mutation, sufficient expression and SDH activity is maintained thanks to the paternal allele. Succinate, ROS and HIF1␣ levels remain below the
threshold and tumor formation is suppressed.
Q7 Figure modified from Müller [52].

330 6.3. The Baysal methylated boundary element model of CpG island methylation, the authors widened their search for 346

a DMR-like element and found two CpG islands (#1 & #2) 9.5 kb 347

331 Another model that addresses the SDHD-linked parent-of-origin upstream of UPGL that function as an alternative promoter. 348

332 dependent inheritance pattern focuses on the idea that expression As a previous study had identified a strong binding site for the 349

333 at or near the SDHD locus determines the pattern of inheritance chromatin regulator CCCTC-binding factor (CTCF) within a UPGL 350

334 [30]. This model is broadly similar to the Muller model in that it CpG island, the authors performed ChIP analyses and confirmed 351

335 supposes differential activity of the two parental alleles at the SDHD strong CTCF binding at CpG island #1 in adrenal and lung tissues 352

336 locus but with an emphasis on the underlying genetic mechanism. [30]. However, in contrast to imprinted loci such as the IGF2/H19 353

337 Baysal and colleagues studied the SDHD region, in which they locus, there was no evidence of significant CpG methylation in the 354

338 identified a differentially methylated CpG island that serves as an CTCF-binding site in either normal tissue or in PGL tumors. Cohesin 355

339 alternative promoter for a large intergenic non-coding (lincRNA), is a chromatin-associated protein complex with a diverse range of 356

340 which the authors termed Untranslated in ParaGanglioma Locus functions, one of which includes co-localizing with CTCF to act as 357

341 (UPGL) [30]. This element is located at the boundary between the an isolator in the IGF2/H19 region. The authors found that RAD21, 358

342 SDHD locus and a flanking gene desert and the authors speculated a cohesin subunit, showed increased binding in fetal adrenal tissue 359

343 that as such it resembles the IGF2/H19 region and therefore might relative to fetal lung and brain [30]. 360

344 also harbor a differentially methylated region (DMR) as controlling Mechanistically, Baysal and colleagues went on to speculate that 361

345 element. As the immediate UPGL promoter showed no evidence CTCF-cohesin binding to the hypermethylated maternal allele of 362

Please cite this article in press as: Hoekstra AS, et al. Models of parent-of-origin tumorigenesis in hereditary paraganglioma. Semin Cell
Dev Biol (2015), http://dx.doi.org/10.1016/j.semcdb.2015.05.011
 G Model
YSCDB 1773 1–8 ARTICLE IN PRESS
6 A.S. Hoekstra et al. / Seminars in Cell & Developmental Biology xxx (2015) xxx–xxx

Fig. 3. Differential long-range accessibility of a downstream enhancer to the SDHD promoter by genomic imprinting. On the paternal allele, a cis-acting element such as
an enhancer may gain access to the SDHD promoter, since cohesion engagement to CTCF may be competitively inhibited by a transcription factor bound to unmethylated
alternative UPGL promoter, resulting in high SDHD expression. On the maternal allele, the alternative UPGL promoter is methylated, making the enhancer instead available
to the UPGL promoter, resulting in low SDHD expression.
Q8 Figure modified from Baysal (2013).

363 the UPGL alternative promoter creates an insulator that interferes absent at the SDHA, SDHB and SDHC loci, explaining the lack of a 396

364 with a functional interaction of the SDHD promoter and a cis-acting parent-of-origin effect on these genes. 397

365 element (enhancer) (Fig. 3). The hypomethylated paternal allele

366 lacks cohesin and thus permits long-range interaction between an 7. Conclusions 398
367 enhancer and the SDHD promoter. This proposed mechanism of
368 differential allelic interaction is hypothesized to result in maternal The models discussed here all attempt to provide viable 399
369 downregulation of SDHD and parent-of-origin-dependent tumor hypotheses that could serve as starting points for experimental 400
370 susceptibility. verification. Although few results have been published to date, 401
371 SDHD has previously been shown to be biallellically expressed the available data are inconclusive and any of these three models 402
372 in fetal brain, kidney and lung [10]. The authors also observed bial- may prove to be correct. Indeed, an entirely unexpected mecha- 403
373 lelic SDHD expression in the adrenal gland (n = 4) and in placenta nism may prove to underlie the striking parent-of-origin effects in 404
374 (n = 3). RT-PCR and sequencing analyses to estimate allelic expres- SDHD/SDHAF2-related PGL/PCC. 405
375 sion of the UPGL gene in seven fetuses showed biallelic expression That said, we currently lean toward the Hensen model. Although 406
376 in the adrenal gland, lung, kidney, skin and brain tissues in six out this model still lacks convincing experimental verification, the 407
377 of seven fetuses. In addition, the authors found a loss of the mater- model is consistent with the available data on paternal inheritance 408
378 nal allele due to loss of heterozygosity [30]. These data appear to and explains: (1) the association of SDHD and SDHAF2, but not 409
379 argue against the mechanism proposed by Baysal and colleagues. In SDHA, SDHB and SDHC, with parent-of-origin dependent tumori- 410
380 addition, a second clear prediction of the Baysal model – that mater- genesis, (2) the somatic loss of maternal chromosome 11, (3) the 411
381 nal inheritance of a mutation would be non-permissive – has now absence of monoallelic expression or DNA methylation at the SDHD 412
382 been challenged by two recent papers that reported paraganglioma locus, and (4) the molecular data from the rare cases with maternal 413
383 development upon maternal transmission of SDHD mutations; both inheritance. A model explaining the inheritance and uniparental 414
384 studies demonstrated LOH at paternal 11q23 and LOH affecting a disomy associated with MAX mutations has not yet been elab- 415
385 section of the maternal chromosome [49,50]. These studies support orated. However, one cannot help but notice the ‘paternal-only’ 416
386 a model in which chromosomal loss of heterozygosity eliminates tumor development shared by SDHD, SDHAF2 and MAX mutations, 417
387 the wild-type SDHD allele (in this case, paternal) together with a genes that are located on chromosomes (11 and 14) with areas of 418
388 loss of a mitotically recombined maternal 11p. The finding that imprinting. These observations suggest that loss of a maternally 419
389 maternally inherited mutations can trigger tumorigenesis (as long expressed gene which is paternally imprinted could be involved in 420
390 as maternal 11p is also lost in tumor) directly contradicts a model MAX-linked PGL/PCC. 421
391 in which the paternal allele is the principle driver of tumorigenesis. In summary, the Hensen model hypothesizes that SDHD-related 422
392 A further issue with the Baysal model is that this elaborate tumorigenesis (in most cases) requires a paternally transmitted 423
393 genetic mechanism would need to be replicated at the SDHAF2 locus mutation, loss of a functional and non-imprinted maternal allele, 424
394 in order to explain the close similarities to SDHD in terms of inher- and loss of a still unidentified maternally expressed, imprinted 425
395 itance, phenotype and penetrance. This mechanism should also be gene(s) elsewhere on chromosome 11. Conceptually, the Hensen 426

Please cite this article in press as: Hoekstra AS, et al. Models of parent-of-origin tumorigenesis in hereditary paraganglioma. Semin Cell
Dev Biol (2015), http://dx.doi.org/10.1016/j.semcdb.2015.05.011
 G Model
YSCDB 1773 1–8 ARTICLE IN PRESS
A.S. Hoekstra et al. / Seminars in Cell & Developmental Biology xxx (2015) xxx–xxx 7

427 model follows the well-known maxim of real estate agents: “loca- oxidative stress, and genomic instability. Cancer Res 2006;66(15): 505

428 tion, location, location”, in that the chromosomal real estate 7615–20. 506
[23] Smith EH, Janknecht R, Maher III LJ. Succinate inhibition of 507
429 occupied by a gene, and not the specific characteristics of the gene alpha-ketoglutarate-dependent enzymes in a yeast model of paraganglioma. 508
430 itself, determines the parent-of-origin effect and perhaps also other Hum Mol Genet 2007;16(24):3136–48. 509
431 aspects of SDHD/SDHAF2 phenotypes. [24] Bayley JP, Kunst HP, Cascon A, Sampietro ML, Gaal J, Korpershoek E, et al. 510
SDHAF2 mutations in familial and sporadic paraganglioma and 511
phaeochromocytoma. Lancet Oncol 2010;11(4):366–72. 512
432 Acknowledgement [25] Cascon A, Robledo M. MAX and MYC: a heritable breakup. Cancer Res 513
2012;72(13):3119–24. 514
[26] Burnichon N, Cascon A, Schiavi F, Morales NP, Comino-Mendez I, Abermil N, 515
Q5
433 This work was supported by the Dutch Cancer Society (Grant et al. MAX mutations cause hereditary and sporadic pheochromocytoma and 516
434 2011-5025 to JPB/PD). paraganglioma. Clin Cancer Res 2012;18(10):2828–37. 517
[27] Riordan JD, Keng VW, Tschida BR, Scheetz TE, Bell JB, Podetz-Pedersen KM, 518
et al. Identification of rtl1, a retrotransposon-derived imprinted gene, as a 519
435 References novel driver of hepatocarcinogenesis. PLoS Genet 2013;9(4):e1003–441. 520
[28] van Baars FM, Cremers CW, van den BP, Veldman JE. Familiar 521

436 [1] Lack E. Tumors of the Adrenal Gland and Extra-Adrenal Paraganglia. AFIP non-chromaffinic paragangliomas (glomus tumors). Clinical and genetic 522

437 Fasicle No. 19. 1997. Washington, DC: American Registry of Pathology; 1997. aspects (abridged). Acta Otolaryngol 1981;91(5–6):589–93. 523

438 [2] Jansen JC, van den BR, Kuiper A, Van Der Mey AG, Zwinderman AH, Cornelisse [29] Mariman EC, van Beersum SE, Cremers CW, Struycken PM, Ropers HH. Fine 524

439 CJ. Estimation of growth rate in patients with head and neck paragangliomas mapping of a putatively imprinted gene for familial non-chromaffin 525

440 influences the treatment proposal. Cancer 2000;88(12):2811–6. paragangliomas to chromosome 11q13.1: evidence for genetic heterogeneity. 526

441 [3] Tischler AS. Pheochromocytoma and extra-adrenal paraganglioma: updates. Hum Genet 1995;95(1):56–62. 527

442 Arch Pathol Lab Med 2008;132(8):1272–84. [30] Baysal BE, McKay SE, Kim YJ, Zhang Z, Alila L, Willett-Brozick JE, et al. 528

443 [4] Cascon A, Comino-Mendez I, Curras-Freixes M, de Cubas AA, Contreras L, Genomic imprinting at a boundary element flanking the SDHD locus. Hum 529

444 Richter S, et al. Whole-exome sequencing identifies MDH2 as a new familial Mol Genet 2011;20(22):4452–61. 530

445 paraganglioma gene. J Natl Cancer Inst 2015;107(5). [31] Andreasson A, Kiss NB, Caramuta S, Sulaiman L, Svahn F, Backdahl M, et al. 531

446 [5] Dahia PL. Pheochromocytoma and paraganglioma pathogenesis: learning The VHL gene is epigenetically inactivated in pheochromocytomas and 532

447 from genetic heterogeneity. Nat Rev Cancer 2014;14(2):108–19. abdominal paragangliomas. Epigenetics 2013;8(12):1347–54. 533

448 [6] Hao HX, Khalimonchuk O, Schraders M, Dephoure N, Bayley JP, Kunst H, et al. [32] Cascon A, Ruiz-Llorente S, Fraga MF, Leton R, Telleria D, Sastre J, et al. Genetic 534

449 SDH5, a gene required for flavination of succinate dehydrogenase, is mutated and epigenetic profile of sporadic pheochromocytomas. J Med Genet 535

450 in paraganglioma. Science 2009;325(5944):1139–42. 2004;41(3):e30. 536

451 [7] Van Der Mey AG, Maaswinkel-Mooy PD, Cornelisse CJ, Schmidt PH, van de [33] Bourgeron T, Rustin P, Chretien D, Birch-Machin M, Bourgeois M, 537

452 Kamp JJ. Genomic imprinting in hereditary glomus tumours: evidence for Viegas-Pequignot E, et al. Mutation of a nuclear succinate dehydrogenase 538

453 new genetic theory. Lancet 1989;2(8675):1291–4. gene results in mitochondrial respiratory chain deficiency. Nat Genet 539

454 [8] Comino-Mendez I, Gracia-Aznarez FJ, Schiavi F, Landa I, Leandro-Garcia LJ, 1995;11(2):144–9. 540

455 Leton R, et al. Exome sequencing identifies MAX mutations as a cause of [34] Piruat JI, Pintado CO, Ortega-Saenz P, Roche M, Lopez-Barneo J. The 541

456 hereditary pheochromocytoma. Nat Genet 2011;43(7):663–7. mitochondrial SDHD gene is required for early embryogenesis, and its partial 542

457 [9] Astuti D, Latif F, Dallol A, Dahia PL, Douglas F, George E, et al. Gene mutations deficiency results in persistent carotid body glomus cell activation with full 543

458 in the succinate dehydrogenase subunit SDHB cause susceptibility to familial responsiveness to hypoxia. Mol Cell Biol 2004;24(24):10933–40. 544

459 pheochromocytoma and to familial paraganglioma. Am J Hum Genet [35] Bayley JP, van M, Hogendoorn I, Cornelisse PC, van der Wal CJ, Prins AFA, et al. 545

460 2001;69(1):49–54. Sdhd and SDHD/H19 knockout mice do not develop paraganglioma or 546

461 [10] Baysal BE, Ferrell RE, Willett-Brozick JE, Lawrence EC, Myssiorek D, Bosch A, pheochromocytoma. PLoS ONE 2009;4(11):e7987. 547

462 et al. Mutations in SDHD, a mitochondrial complex II gene, in hereditary [36] Dannenberg H, de Krijger RR, Zhao J, Speel EJ, Saremaslani P, Dinjens WN, 548

463 paraganglioma. Science 2000;287(5454):848–51. et al. Differential loss of chromosome 11q in familial and sporadic 549

464 [11] Burnichon N, Briere JJ, Libe R, Vescovo L, Riviere J, Tissier F, et al. SDHA is a parasympathetic paragangliomas detected by comparative genomic 550

465 tumor suppressor gene causing paraganglioma. Hum Mol Genet hybridization. Am J Pathol 2001;158(6):1937–42. 551

466 2010;19(15):3011–20. [37] Hensen EF, Jordanova ES, van Minderhout IJHM, Hogendoorn PCW, Taschner 552

467 [12] Niemann S, Muller U. Mutations in SDHC cause autosomal dominant PEM, van der Mey AGL, et al. Somatic loss of maternal chromosome 11 causes 553

468 paraganglioma, type 3. Nat Genet 2000;26(3):268–70. parent-of-origin-dependent inheritance in SDHD-linked paraganglioma and 554

469 [13] Douwes Dekker PB, Hogendoorn PC, Kuipers-Dijkshoorn N, Prins FA, van phaeochromocytoma families. Oncogene 2004;23(23):4076–83. 555

470 Duinen SG, Taschner PE, et al. SDHD mutations in head and neck [38] Riemann K, Sotlar K, Kupka S, Braun S, Zenner HP, Preyer S, et al. Chromosome 556

471 paragangliomas result in destabilization of complex II in the mitochondrial 11 monosomy in conjunction with a mutated SDHD initiation codon in 557

472 respiratory chain with loss of enzymatic activity and abnormal mitochondrial nonfamilial paraganglioma cases. Cancer Genet Cytogenet 558

473 morphology. J Pathol 2003;201(3):480–6. 2004;150(2):128–35. 559

474 [14] Koivunen P, Hirsila M, Remes AM, Hassinen IE, Kivirikko KI, Myllyharju J. [39] Beristain E, Vicente MA, Guerra I, Gutierrez-Corres FB, Garin I, Perez de NG. 560

475 Inhibition of hypoxia-inducible factor (HIF) hydroxylases by citric acid cycle Disomy as the genetic underlying mechanisms of loss of heterozigosity in 561

476 intermediates – possible links between cell metabolism and stabilization of SDHD-paragangliomas. J Clin Endocrinol Metab 2013;98(5)):E1012–6. 562

477 HIF. J Biol Chem 2007;282(7):4524–32. [40] Eggermann T, Binder G, Brioude F, Maher ER, Lapunzina P, Cubellis MV, et al. 563

478 [15] MacKenzie ED, Selak MA, Tennant DA, Payne LJ, Crosby S, Frederiksen CM, CDKN1C mutations: two sides of the same coin. Trends Mol Med 564

479 et al. Cell-permeating alpha-ketoglutarate derivatives alleviate 2014;20(11):614–22. 565

480 pseudohypoxia in succinate dehydrogenase-deficient cells. Mol Cell Biol [41] Castro-Vega LJ, Letouze E, Burnichon N, Buffet A, Disderot PH, Khalifa E, et al. 566

481 2007;27(9):3282–9. Multi-omics analysis defines core genomic alterations in 567

482 [16] Selak MA, Armour SM, MacKenzie ED, Boulahbel H, Watson DG, Mansfield KD, pheochromocytomas and paragangliomas. Nat Commun 2015;6:6044. 568

483 et al. Succinate links TCA cycle dysfunction to oncogenesis by inhibiting [42] Corver WE, Middeldorp A, ter Haar NT, Jordanova ES, van PM, van ER, et al. 569

484 HIF-alpha prolyl hydroxylase. Cancer Cell 2005;7(1):77–85. Genome-wide allelic state analysis on flow-sorted tumor fractions provides 570

485 [17] Albayrak T, Scherhammer V, Schoenfeld N, Braziulis E, Mund T, Bauer MKA, an accurate measure of chromosomal aberrations. Cancer Res 571

486 et al. The tumor suppressor cybL, a component of the respiratory chain, 2008;68(24):10333–40. 572

487 mediates apoptosis induction. Mol Biol Cell 2003;14(8):3082–96. [43] Margetts CDE, Astuti D, Gentle DC, Cooper WN, Cascon A, Catchpoole D, et al. 573

488 [18] Goffrini P, Ercolino T, Panizza E, Giache V, Cavone L, Chiarugi A, et al. Epigenetic analysis of HIC1, CASP8, FLIP, TSP1, DCR1, DCR2, DR4, DR5, 574

489 Functional study in a yeast model of a novel succinate dehydrogenase subunit KvDMR1, H19 and preferential 11p15.5 maternal-allele loss in von 575

490 B gene germline missense mutation (C191Y) diagnosed in a patient affected Hippel–Lindau and sporadic phaeochromocytomas. Endocr Relat Cancer 576

491 by a glomus tumor. Hum Mol Genet 2009;18(10):1860–8. 2005;12(1):161–72. 577

492 [19] Guzy RD, Sharma B, Bell E, Chandel NS, Schumacker PT. Loss of the SdhB, but [44] Hensen EF, Jansen JC, Siemers MD, Oosterwijk JC, Vriends AH, Corssmit EP, 578

493 not the SdhA, subunit of complex II triggers reactive oxygen et al. The Dutch founder mutation SDHD.D92Y shows a reduced penetrance 579

494 species-dependent hypoxia-inducible factor activation and tumorigenesis. for the development of paragangliomas in a large multigenerational family. 580

495 Mol Cell Biol 2008;28(2):718–31. Eur J Hum Genet 2010;18(1):62–6. 581

496 [20] Ishii T, Yasuda K, Akatsuka A, Hino O, Hartman PS, Ishii N. A mutation in the [45] Kunst HP, Rutten MH, De Monnink JP, Hoefsloot LH, Timmers HJ, Marres HA, 582

497 SDHC gene of complex II increases oxidative stress, resulting in apoptosis and et al. SDHAF2 (PGL2-SDH5) and hereditary head and neck paraganglioma. Clin 583

498 tumorigenesis. Cancer Res 2005;65(1):203–9. Cancer Res 2011;17(2):247–54. 584

499 [21] Owens KM, Aykin-Burns N, Dayal D, Coleman MC, Domann FE, Spitz DR. [46] Hes FJ, Weiss MM, Woortman SA, de Miranda NF, van Bunderen PA, Bonsing 585

500 Genomic instability induced by mutant succinate dehydrogenase subunit D BA, et al. Low penetrance of a SDHB mutation in a large Dutch paraganglioma 586

Q6
501 (SDHD) is mediated by O2 (−• ) and H2 O2 . Free Radic Biol Med family. BMC Med Genet 2010;11:92. 587

502 2012;52(1):160–6. [47] Schiavi F, Milne RL, Anda E, Blay P, Castellano M, Opocher G, et al. Are we 588

503 [22] Slane BG, ykin-Burns N, Smith BJ, Kalen AL, Goswami PC, Domann FE, et al. overestimating the penetrance of mutations in SDHB? Hum Mutat 589

504 Mutation of succinate dehydrogenase subunit C results in increased O2 • −, 2010;31(6):761–2. 590

Please cite this article in press as: Hoekstra AS, et al. Models of parent-of-origin tumorigenesis in hereditary paraganglioma. Semin Cell
Dev Biol (2015), http://dx.doi.org/10.1016/j.semcdb.2015.05.011
 G Model
YSCDB 1773 1–8 ARTICLE IN PRESS
8 A.S. Hoekstra et al. / Seminars in Cell & Developmental Biology xxx (2015) xxx–xxx

591 [48] Solis DC, Burnichon N, Timmers HJ, Raygada MJ, Kozupa A, Merino MJ, et al. [52] Muller U. Pathological mechanisms and parent-of-origin effects in hereditary 603
592 Penetrance and clinical consequences of a gross SDHB deletion in a large paraganglioma/pheochromocytoma (PGL/PCC). Neurogenetics 604
593 family. Clin Genet 2009;75(4):354–63. 2011;12(3):175–81. 605
594 [49] Yeap PM, Tobias ES, Mavraki E, Fletcher A, Bradshaw N, Freel EM, et al. [53] Cervera AM, Bayley JP, Devilee P, McCreath KJ. Inhibition of succinate 606
595 Molecular analysis of pheochromocytoma after maternal transmission of dehydrogenase dysregulates histone modification in mammalian cells. Mol 607
596 SDHD mutation elucidates mechanism of parent-of-origin effect. J Clin Cancer 2009;8:89. 608
597 Endocrinol Metab 2011;96(12):E2009–13. [54] Letouze E, Martinelli C, Loriot C, Burnichon N, Abermil N, Ottolenghi C, et al. 609
598 [50] Bayley JP, Oldenburg RA, Nuk J, Hoekstra AS, van der Meer CA, Korpershoek E, SDH mutations establish a hypermethylator phenotype in paraganglioma. 610
599 et al. Paraganglioma and pheochromocytoma upon maternal transmission of Cancer Cell 2013;23(6):739–52. 611
600 SDHD mutations. BMC Med Genet 2014;15:111. [55] Lee S, Nakamura E, Yang H, Wei W, Linggi MS, Sajan MP, et al. Neuronal 612
601 [51] Niemann S, Steinberger D, Muller U. PGL3, a third, not maternally imprinted apoptosis linked to EglN3 prolyl hydroxylase and familial pheochromocytoma 613
602 locus in autosomal dominant paraganglioma. Neurogenetics genes: developmental culling and cancer. Cancer Cell 2005;8(2): 614
1999;2(3):167–70. 155–67. 615

Please cite this article in press as: Hoekstra AS, et al. Models of parent-of-origin tumorigenesis in hereditary paraganglioma. Semin Cell
Dev Biol (2015), http://dx.doi.org/10.1016/j.semcdb.2015.05.011