The 3PAs: An Update on the Association of Pheochromocytomas, Paragangliomas, and Pituitary Tumors

Abstract

Pituitary adenomas (PA) and pheochromocytomas/paragangliomas (PHEO/PGL) are rare tumors. Although they may co-exist by coincidence, there is mounting evidence that genes predisposing in PHEO/PGL development, may play a role in pituitary tumorigenesis. In 2012, we described a GH-secreting PA caused by an SDHD mutation in a patient with familial PGLs and found loss of heterozygosity at the SDHD locus in the pituitary tumor, along with increased hypoxia-inducible factor 1α (HIF-1α) levels. Additional patients with PAs and SDHx defects have since been reported. Overall, prevalence of SDHx mutations in PA is very rare (0.3–1.8 % in unselected cases) but we and others have identified several cases of PAs with PHEOs/PGLs, like our original report, a condition which we termed the 3 P association (3PAs). Interestingly, when 3PAs is found in the sporadic setting, no SDHx defects were identified, whereas in familial PGLs, SDHx mutations were identified in 62.5–75 % of the reported cases. Hence, pituitary surveillance is recommended among patients with SDHx defects. It is possible that the SDHx germline mutation-negative 3PAs cases may be due to another gene, epigenetic changes, mutations in modifier genes, mosaicism, somatic mutations, pituitary hyperplasia due to ectopic hypothalamic hormone secretion or a coincidence. PA in 3PAs are mainly macroadenomas, more aggressive, more resistant to somatostatin analogues, and often require surgery. Using the Sdhb ^{+ / −} mouse model, we showed that hyperplasia may be the first abnormality in tumorigenesis as initial response to pseudohypoxia. We also propose surveillance and follow-up approach of patients presenting with this association.

Introduction

Pheochromocytomas and paragangliomas (PHEOs/PGLs) are rare neuroendocrine tumors that produce catecholamines and arise from three structures derived from the neural crest: the adrenal medulla (PHEOs) and the sympathetic and parasympathetic paraganglia (PGLs) [1]. Most PHEOs are sporadic, unicentric and unilateral, but up to half (or even more, depending on the age of presentation) may be familial, multicentric and bilateral [2]. In recent years, substantial progress has been accomplished in the field of genetics and pathogenesis of PHEO/PGL and we know now that there are more than 20 susceptibility genes [3] including the genes encoding the four succinate dehydrogenase complex (SDH) subunits (SDHA, -B, -C, -D or SDHx collectively) [4–7]. Mutations/functional variants of SDHx have also been implicated in Carney–Stratakis dyad or syndrome and, rarely, in Carney triad [8], renal cancer [9–12], pancreatic neuroendocrine tumors [13] and Cowden-like syndrome [14].

Frequency of SDHx mutations in pituitary adenomas and the new syndromic association (3PAs)

In 2012, we described a family with multiple PGLs and PHEOs caused by a germline SDHD mutation; the index case also had an aggressive GH-secreting pituitary adenoma (PA) [15]. We found loss of heterozygosity (LOH) at the SDHD locus in the pituitary tumor, along with increased hypoxia-inducible factor 1α (HIF-1α) levels. These findings indicated that the SDH defect was most likely causatively linked to the development of the pituitary tumor and that pseudohypoxia pathways were activated, as shown in PHEOs/PGLs that bear SDHx mutations [15, 16].

However, what was also significant, was that the case of this patient was similar to several other cases previously reported since 1952; in most of these reports, the co-existence of PHEOs/PGLs and PAs was thought to represent a mere coincidence [17]. Following our 2012 publication, additional cases of PAs among patients with SDHx defects were described; today, it is widely accepted that the association of PAs with PHEOs and/or PGLs, or the 3 P association (3PAS) represents a new inherited form of a predisposition to multiple endocrine tumors caused by SDHx defects [18–25].

Indeed, to investigate the frequency of germline SDHx mutations in PAs we sequenced 168 patients with sporadic and familial PAs. Overall, SDHx mutations were rare: 1.8 % of the studied cases [26], in accordance with other reports [18, 25, 27]. However, among the patients included in the cohort, there were 7 that had presented with medical or family history of PHEOs/PGLs [26]; these patients had SDHx sequence pathogenic defects. Whenever family samples were available, we were able to show that SDHx defects were associated with either PHEOs/PGLs or PAs or both. On the other hand, in the few patients where PHEOs/PGLs and PAs were found in the sporadic setting (there was no family history or prior medical history of other SDHx-related defects), no SDHx mutations were identified. The difference was highly significant with more than 75 % of patients with 3PAs and positive family history bearing SDHx mutations [26]. The overall prevalence of 3PAs with SDHx mutations irrespective of family history was 42.8 % (3 out of 7 detected cases). Similar results were obtained in a different cohort of patients with PAs. Denes et al. studied 39 patients with sporadic or familial PHEO/PGL and/or PAs. Eight patients with SDHx mutations or variants within an international cohort of 19 patients with both tumors, were identified, which accounted for 42.1 % of the affected subjects. Five out of the 8 cases (62.5 %) were found within families with familial PHEOs/PGLs [22]. There was also a single patient with an SDHAF2 variant located in the 5′-UTR.

Interestingly, in the cohort studied by Denes et al., there were also 4 cases who presented with PHEO/PGLs and PAs but were found to harbor VHL or MEN1 gene mutations [22]. These 2 cases of PA and PHEO/PGL co-existence in patients with VHL mutations are the only ones reported in the literature, so far. Considering the frequency with which patients with VHL undergo regular surveillance imaging of the brain, the low frequency may indicate that the association of VHL and PA probably does not represent a true (genetically linked) predisposition. On the other hand in the 2 cases with MEN1 mutations, LOH was identified in the available PHEO tissue, indicating that PHEOs and/or PGLs can be part of the MEN 1 syndrome (as the Men1 knock-out animal model also suggests); thus, genetic testing for menin mutations should be considered in patients with PHEO/PGL if there are other suggestive signs of MEN 1. MEN 4 (the CDKN1B gene, coding for p27) may also be screened for in patients with an MEN 1-like phenotype, PAs, and PHEOs/PGLs that do not have menin or SDHx defects [22, 28–32].

We reviewed all the reported cases so far of the combination of PAs and PHEOs/PGLs since 1952 (eighty-two in total), including a recent report published while our manuscript was under review [33]. Literature search was done using PubMed, Scopus, and Google Scholar as search engines and “pituitary adenoma”, “3PAs”, “pheochromocytoma”, “paraganglioma” and combinations as search terms. Thirty-one (37.80 %) of these cases harbored mutations in predisposing PHEOs/PGLs or PA genes. Twenty-two patients (26.82 %) of all 3PAs cases had a personal or family history suggestive of a hereditary endocrine syndrome, whereas thirty-seven of all cases (45.1 %) were isolated; for the rest 28 % no information regarding family history was available (Table 1, 2). Among the 82 3PAs cases described so far, 17 % had both identified genetic mutations and family history. This co-existence was much higher (48.38 %) if we consider only the cases with a genetic defect. Regarding the frequency of the identified genetic defects it is obvious that the majority of cases carry SDHx defects (19 out of 31 cases, 61.3 %), with MEN1 and MAX being the second and third most frequently related genes. Of course, due to the retrospective nature of this review and the lack of genetic screening in the majority of 3PAs cases, generalizations regarding the overall frequency of the implicated genes should be avoided.

Table 1.

Patients with pituitary adenoma and pheochromocytoma/paraganglioma syndrome (3PAs) with identified germline genetic defects.

		Pituitary				Pheo/PGL
Patient Nr	Sex	Type	Size	Treatment	Age	Type	Treatment	Age	Family history	Mutation	LOH/ICH in PA	Reference
1	F	PRL	NK	NK	27	Pheo	NK	NK	No	SDHA c.91C > T p.Arg31Ter, VHL c.589G > A p.Asp197Asn	Not performed	Dénes et al. (2015) [22]
2	F	PRL	Macro	NK	49	PGLs	NK	49	NK	SDHA p.Arg31 * c.91C > T	SDHA/SDHB negative staining	Niemeijer et al. (2015) [13]
3	M	GH	Macro	SSA	84	PGL	No	84	No	SDHAF2 c.−52T > C	Not performed	Dénes et al. (2015) [22]
4	M	PRL	Macro	DA, surgery	33	PGL	Surgery	33	Mother: PRL, brother: PGL	SDHB c.298 T > C p.Ser100Pro	LOH at SDBH locus, intracytoplasmic vacuoles	Dénes et al. (2015) [22]
5	F	NFPA	Macro	Surgery x3, RT	53	PGL	RT	28	Sister: glioma	SDHB c.587 G > A p.Cys196Tyr	LOH at SDBH locus/SDHB staining: diffuse/intracytoplasmic vacuoles	Dénes et al. (2015) [22]
6	F	PRL	Macro	DA	38	PGLs carotid and mediastinal	Carotid: surgery, mediastinal: inoperable	35	Brother PGLs	SDHB mutation, (actual genetic defect not available)	Not performed	Gorospe et al. (2017) [76]
7	F	PRL	Macro	DA, RT	60	PGL	RT	60	NK	SDHB c.423 + 1 G > A	Not performed	Dénes et al. (2015) [22]
8	F	NFPA	Micro	No	50	Pheo	Surgery	50	NK	SDHB c.770dupT p.Asn-258GlufsTer17	Not performed	Dénes et al. (2015) [22]
9	M	GH	NK	SSA	72	PGL	No	70	Brother & niece: PA, sister: bilateral HNPGL	SDHB c.689 G > A p.Arg230His	Not performed	Xekouki et al. (2015) [26]
10	F	PRL	Micro	NK	50	PGL	NK	47	Brother: HNPGL, grandmother: GIST	SDHB c.642 + 1 G > A, splice site alteration	Not performed	Xekouki et al. (2015) [26]
11	F	PRL	Micro	DA	33	PGL	Surgery	43	Brain tumor	SDHB c.18 C > A p.Ala6Alaa 3 PTEN polymorphisms	Not performed	Efstathiadou et al. (2014) [77]
12	F	PRL	Macro	DA	38	PGL	SSA	NK	Brother index case: PGL. Mother and sister positive for region of ex.1 of SDHB	deletion affecting ex. 1 of SDHB	Not performed	Guerrero Pérez et al. (2016) [21]
13	M	PRL	Macro	DA	53	PGL	Surgery	38	Cousin: PA, brother: PGL	SDHC c.380 A > G p.His127Arg	Not performed	Dénes et al. (2015) [22]
14	F	PRL	Macro	NK	60	PGL	NK	60	No	SDHC c256–257insTTTp-Phe85dup	Not performed	López-Jiménez et al. (2008) [78]
15	F	PRL	Macro	Surgery, DA	23	PGL	Surgery	32	Sister, aunt and grandmother: PA; sister: bilateral HNPGL	SDHD c.242 C > T, p.Pro81Leu	Not performed	Xekouki et al. (2015) [26]
16	M	PRL	Macro	DA, surgery	60	PGL, Pheo	Surgery (Pheo)	62	NK	SDHD c.274 G > T pAsp92Tyr	LOH at SDHD locus/SDHB positive ICH, SDHA IHC positive	Papathomas et al. (2014) [25]
17	F	GH	Macro	Surgery, SSA	56	PGL	NK	56	Father and 2 sisters: HNPGL; sister: GIST	SDHD c.274 G > T p.Asp92Tyr	LOH at SDHD locus/SDHB positive ICH, SDHA IHC positive	Papathomas et al. (2014) [25]
18	F	PRL	Macro	DA, surgery	33	PGL	Surgery x2	39	Aunt, uncle, brother: HNPGL	SDHD c.242 C > T p.Pro81Leu	Not performed	Varsavsky et al. (2013) [79]
19	M	GH	Macro	SSA, surgery	37	PGL, Pheo	Surgery	37	Sister and paternal uncle neck PGLs HNPGL	SDHD c.298_301 del, premature stop at codon 133 AIP & CDKN1B polymorphism	PA: LOH at SDHD locus, reduced SDHD protein/patchy SDHB staining	Xekouki et al. (2012) [15,17]
20	M	GH/PRL	Macro	Surgery, RT, DA	27	Pheo	Surgery	31	No	MEN1 c.1452delGp.Thr557Ter	Menin staining of the Pheo: no menin positive cells	Dénes et al. (2015) [22]
21	F	NFPA	Macro	Surveillance	45	PGL	NK	45	No	MEN1 c.196_200dupAG-CCC frameshift (pathogenic), polymorphism C423T no amino acid change	Not performed in Pheo	Jeong et al. (2014) [80]
22	M	PRL	NK	NK	41	Pheo	Surgery	48	NK	MEN1 K119X, RET WT	Not performed in Pheo	Langer et al. (2002) [81]
23	NK	NK	NK	NK	NK	Pheo	NK	NK	Hyperparathy-roidism and pancreatic islet cell tumor	MEN1 c.320dell2	Not performed in Pheo	Dackiw et al. (1999) [32]
24	NK	NK	NK	NK	NK	Pheo	NK	NK	Pancreatic islet cell tumor and rectal leiomyoma	MEN1 1325insA	Not performed in Pheo	Dackiw et al. (1999) [32]
25	M	PRL	NK	NK	29	Pheo	NK	32	MEN1	Not performed, other family members MEN1 mutation	HPTH at age 21 years	Carty et al. (1998) [82]
26	M	GH	Macro	Surgery	62	Pheo	Surgery	62	No	RET p.Cys618Ser	Not performed in PA	Heinlen et al. (2011) [83]
27	M	ACTH	Micro	Surgery x2	48	Pheo	Surgery	66	Son: HPTH	RET c.1900T > C, p.Cys634Arg. Negative for MEN1 mutations	Not performed in Pheo	Naziat et al. (2013) [84]
28	F	PRL	Macro	DA	49	Pheo	Bilateral adrenalectomy	49	No	MAX c.296 G > T, Neg for MEN1, VHL, SDHB, SDHC, SDHD, SDHAF2, or TMEM127 genes	Not performed	Roszko et al. (2017) [34]
29	M	PRL	Micro	DA	49	Pheo	Surgery	32	No	Germinal heterozygous deletion of exon 3 of MAX (detected by MLPA). Neg for RET, VHL, SDHx, CDKN1B, AIP, MEN1	Not performed	Daly et al., (2018) [33]
30	F	GH	Macro	SSA, DA, Pegvisomant, RT	26	Bilateral Pheos	Bilateral adrenalectomy	35	No	Germinal heterozygous deletion of exons 1–3 and intron 3 of MAX. Neg for RET, VHL, SDHx, CDKN1B, AIP, MEN1	Not performed	Daly et al. (2018) [33]
31	M	GH	Macro	Surgery, RT	16	Bilateral Pheos	Bilateral adrenalectomy	22	No	Germinal heterozygous deletion of exon 3 of MAX (detected by MLPA). Neg for RET, VHL, SDHx, CDKN1B, AIP, MEN1	Not performed	Daly et al. (2018) [33]

M: Male; F: Female; NFPA: Non-functional pituitary adenoma; PRL: Prolactinoma; GH: Acromegaly; Macro: Macroadenoma; Micro: Microadenoma; DA: Dopamine agonist; RT: Radiotherapy; SSA: Somatostatin analogue; Pheo: Pheochromocytoma; PGL: Paraganglioma; HNPGL: Head and neck paraganglioma; PTC: Papillary thyroid cancer; GIST: Gastrointestinal stromal tumor; pNET: Pancreatic neuroendocrine tumor; MTC: Medullary thyroid carcinoma; HPTH: Hyperparathyoidism; NK: Not known; MEN1: Multiple endocrine neoplasia type 1; NF1: Neurofibromatosis type 1; A: Single nucleotide polymorphism with a minor allele frequency of 0.2% and a genotype frequency of 0.5% (1000 Genomes Project Consortium, 2012) [85].

Table 2.

Patients with pituitary adenoma and pheochromocytoma/paraganglioma without identified germline genetic defects.

		Pituitary				Pheo/PGL
Patient Nr	Sex	Type	Size	Treatment	Age	Type	Treatment	Age	Family history	Mutation	Other Info (when reported)	Reference
1	M	PRL	Micro	Surgery	54	Pheo	Bilateral adrenalectomy	NK	No	Neg for RET, VHL, SDHB, SDHD mutations		Guerrero Pérez et al. (2016) [21]
2	F	GH	Micro	Surgery	56	Pheo	Bilateral adrenalectomy	NK	No	Neg for RET, VHL, MEN-1 mutations	Bilateral nodules on parathyroid glands, asymptomatic	Guerrero Pérez et al. (2016) [21]
3	M	PRL	Macro	Cabergoline	38	Pheo/PGL	α-Adrenergic blockade	NK	NK	Neg for MEN1	Catecholamine-mediated cardiac toxicity	Koshy et al. (2016) [86]
4	F	ACTH	Micro	Surgery	61	PGL	Surveillance	61	No	Negative for SDHA-D, MEN1, RET, AIP	Bilateral HNPGL	Xekouki et al. (2015) [26]
5	F	PRL	Macro	DA, surgery	35	Pheo	Surgery	55	No	Negative for SDHA-D, MEN1, RET, AIP	Bilateral Pheo	Xekouki et al. (2015) [26]
6	F	GH	Macro	Surgery	35	PGL	Surgery	58	No	Negative for SDHA-D, MEN1, RET, AIP	Bladder PGL	Xekouki et al. (2015) [26]
7	F	NFPA	Macro	Surgery	39	Pheo	Surgery	34	No	Negative for SDHA-D, MEN1, RET, AIP		Xekouki et al. (2015) [26]
8	F	GH	Macro	Surgery, RT, DA, SSA	56	Pheo	Surgery	66	No	Negative for SDHA-D, AF2, MEN1, RET, AIP, VHL, TMEM127, MAX, FH, CDKN1B	GIST, thyroid follicular adenoma	Boguszewski et al. (2012) [87], Dénes et al. (2015) [22]
9	M	NFPA	Macro	Surgery	53	PGL	Surgery	50	Father: PA	SDHA c.969 C > T p.Cly323GlyaSDHB-D, AF2, MEN1, RET, AIP, VHL, TMEM127, MAX, FH, CDKN1B all normal	Abdominal PGL Wilms tumor, liposarcoma, renal oncocytoma; PA: no LOH at SDHA locus, intracytoplasmic vacuoles, SDHA and B staining preserved	Dénes et al. (2015) [22]
10	F	GH	Macro	Surgery, RT, SSA	39	Pheo	Surgery	20	No	Negative for SDHA-D, AF2, MEN1, RET, AIP, VHL, TMEM127, MAX, FH, CDKN1B		Dénes et al. (2015) [22]
11	F	NFPA	Macro	Surgery, RT	73	PGL	RT	73	No	Negative for SDHA-D, AF2, MEN1, RET, AIP, VHL, TMEMU7, MAX, FH, CDKN1B	HNPGL	Dénes et al. (2015) [22]
12	M	GH	Macro	Infarcted	16	Pheo	NK	16	No	Negative for SDHA-D, AF2, MEN1, RET, AIP, VHL, TMEM127, MAX, FH, CDKN1B		Dénes et al. (2015) [22]
13	M	PRL	Macro	Surgery	40 s	PGL	NK	52	No	Negative for SDHA-D, AF2, MEN1, RET, AIR VHL, TMEM127, MAX, FH, CDKN1B	HNPGL	Dénes et al. (2015) [22]
14	F	PRL	NK	NK	27	Pheo	NK	41	No	Negative for SDHA-D, AF2, MEN1, RET, AIP, VHL, TMEM127, MAX, FH, CDKN1B		Dénes et al. (2015) [22]
15	M	NK	NK	NK	NK	Pheo/PGL	NK	NK	No	Negative for SDHA-D, AF2, MEN1, RET, AIP, VHL, TMEM127, MAX, FH, CDKN1B		Denes etal. (2015) [22]
16	F	PRL	Micro	DA	40	Pheo	Surgery	38	No	Negative for SDHA-D, AF2, MEN1, RET, AIP, VHL, TMEM127, MAX, FH, CDKNW		Dénes et al. (2015) [22]
17	M	PRL	Micro	DA	56	Pheo	Surgery	56	No	Negative for SDHA-D, AF2, MEN1, RET, AIP, VHL, TMEM127, MAX, FH, CDKN1B		Dénes et al. (2015) [22]
18	F	PRL	Macro	DA	61	Pheo	Surgery	61	No	Negative for SDHA-D, AF2, MEN1, RET, AIP, VHL, TMEM127, MAX, FH, CDKN1B		Dénes et al. (2015) [22]
19	F	PRL	Macro	NK	60	PGL	RT	60	No	Negative for SDHB		Parghane et al. (2014) [88]
20	F	NFPA	Macro	No	52	Pheo	Surgery	52	No	Negative for SDHA-D, AF2, RET, MAX, TMEMT27, VHL	GHRH secreting Pheo	Mumby et al. (2014) [89]
21	M	GH	Macro	Surgery	29	Pheo	Surgery	29	No	Not performed	Bilateral Pheo lipoma, metastatic PTC	Sisson et al. (2012) [90]
22	NK	GH	NK	NK	NK	Pheo	NK	NK	MEN1	Not performed	Bilateral Pheo HPTH, pNET Clinical features NF1	Gatta-Cherifi et al. (2012) [91]
23	M	GH	Macro	Surgery	45	PGL, Pheo	Surgery × 3	54	Father HNPGL, Sister: adrenal abnormality	Not performed	Abdominal, HN, cardiac PGLs	Zhang et al. (2011) [92]
24	M	NFPA	Micro	No	64	Pheo	Surgery	64	No	Not performed	High cortisol (cured post-adrenalectomy)	Yaylali et al. (2008) [93]
25	M	NFPA	Macro	Surgery	59	Pheo	Surgery	59	No	Not performed		Breckenridge et al. (2003) [94]
26	M	NFPA	Macro	Surgery	56	Pheo	No	56	No	Not performed		Dünser et al. (2002) [95]
27	F	GH	Macro	Surgery	57	Pheo	Surgery	57	No	Not performed		Sleilati et al. (2002) [96]
28	M	NK	Micro	No	43	Pheo	Surgery	43	No	Negative for RET	Lipoma, pectus excavatum, pleomor-phic parotid adenoma. GH levels responded to OGTT post adrenalectomy	Baughan et al. (2001) [97]
29	F	NK	Micro	No	44	Pheo	Surgery	44	No	Not performed	Cushing’s (cured post-adrenalectomy)	Khalil et al. (1999) [98]
30	M	GH	Macro	Surgery	41	Pheo, PGLs	Surgery	20	NK	Negative for RET		Teh et al. (1996) [99]
31	M	PRL	Macro	Surgery	20	PGL	Surgery	20	NK	Not performed	HNPGL	Azzarelli et al. (1988) [100]
32	M	PRL	Macro	DA	26	Pheo	Surgery	26	Father: metastatic MTC and probably Pheo	Not performed		Bertrand et al. (1987) [101]
33	F	NFPA	NK	NK	70	PGL	NK	70	Daughter and granddaughter: PA, bilateral HNPGL	Not performed	HPTH, PTC, gastric leiomyoma, amyloidosis	Larraza-Hernandez et al. (1982) [102]
34	F	PRL	Micro	NK	35	Pheo	NK	35	NK	Not performed		Meyers (1982) [103]
35	M	NFPA	Macro	Surgery	66	PGL	Surgery	63	NK	Not performed		Blumenkopf & Boekelheide (1982) [104]
36	F	GH	NK	Surgery, RT	53	Pheo	Surgery	58	Brother: hypertension	Not performed		Anderson et al. (1981) [105]
37	F	GH	NK	RT	33	Pheo	None	45	No	Not performed	Multinodular goiter	Anderson et al. (1981) [105]
38	F	GH	Macro	NK	53	Pheo	NK	53	NK	Not performed	HPTH	Myers & Eversman (1981) [106]
39	F	PRL	NK	NK	23	Pheo	NK	23	NK	Not performed	HPTH, gastrinoma, adrenal adenoma	Alberts et al. (1980) [107]
40	F	NK	Macro	NK	22	Pheo	NK	22	Granddaughter: unilateral Pheo	Not performed	Islet cell tumor/ renal adenoma	Janson et al. (1978) [108]
41	F	GH	Macro	NK	15	Pheo	NK	15	NK	Not performed	HPTH	Manger & Clifford (1977) [109]
42	F	NFPA	NK	NK	49	Pheo	None	49	NK	Not performed	Papillary carcinoma of thyroid	Melicow (1977) [110]
43	M	GH	NK	NK	21	Pheo	NK	44	NK	Not performed		Kadowaki et al. (1976) [111]
44	F	GH	NK	NK	19	PGL	NK	19	NK	Not performed	PGL (HN, pelvis), HPTH	Farhi etal. (1976) [112]
45	F	GH	NK	NK	36	PGL	NK	36	NK	Not performed	HNPGL, HPTH, hyperplasia of antral and duodenal gastrin cells	Berg et al. (1976) [113]
46	F	NK	Micro	NK	43	Pheo	NK	43	NK	Not performed	MTC	Wolf et al. (1972) [114]
47	F	GH	NK	RT	36	Pheo	Surgery	36	NK	Not performed	toxic nodular, goiter, endometriosis, and diabetes, DM	Miller &Wynn (1971) [115]
48	M	ACTH	NK	NK	NK	Pheo	Surgery	41	VI generations with MEN	Not performed	MTC	Steiner et al. (1968) [116]
49	M	GH	Macro	RT	23	Pheo	None	41	NK	Not performed		Kahn&Mullon (1964) [117]
50	M	GH	NK	NK	NK	Pheo	NK	NK	NK	Not performed		German & Flanigan (1964) [118]
51	M	GH	NK	NK	44	Pheo	NK	NK	NK	Not performed		Iversen (1952) [119]

M: Male; F: Female; NFPA: Non-functional pituitary adenoma; PRL: Prolactinoma; GH: Acromegaly; Macro: Macroadenoma; Micro: Microadenoma; DA: Dopamine agonist; RT: Radiotherapy; SSA: Somatostatin analogue; Pheo: Pheochromocytoma; PGL: Paraganglioma; HNPGL: Head and neck paraganglioma; PTC: Papillary thyroid cancer; GIST: Gastrointestinal stromal tumor; pNET: Pancreatic neuroendocrine tumor; MTC: Medullary thyroid carcinoma; HPTH: Hyperparathyoidism; NK: Not known; MEN1: Multiple endocrine neoplasia type 1; NF1: Neurofibromatosis type 1; GOTT: Glucose oral tolerance test; a Single nucleotide polymorphism with a frequency of 3.5% (Bayley et al. 2005) [120].

In a most recent publication Daly et al., reported 3 patients who presented with 3PAs and were found to have MAX exon/intragenic deletions using multiplex ligation-dependent probe amplification (MLPA), confirming a previous report of Roszko et al. [34]. The authors recommend that MAX MLPA should be considered in 3PAs and in PHEO cases in individuals screened negative with Sanger sequencing for the reported genetic causes including MAX mutations.

Despite the new identified genetic mutations/variants associated with 3PAs, for the majority of cases we do not know what other PHEOs/PGLs-causing genetic defects may be associated with a predisposition to PAs. At this point, what one screens for (beyond SDHx, menin, and possibly p27) remains unknown but it should be guided by detailed medical and family history, the latter extending to previous generations and even distant relatives, as these mutations often have weak penetrance. One should also avoid the effort to “fit” unusual cases into the known conditions: it is evident that the known classification of MEN syndromes does not always cover rare individual cases that present with significant overlap in their phenotype or, even, seemingly unrelated clinical signs that may represent new associations. Regarding the 51 cases of 3PAs with no known genetic defect (Table 2) most of them are older and genetic testing was not available. Therefore, we cannot exclude the possibility that they harbor a genetic defect in any of the implicated genes; for some of them and based on the clinical presentation and family history the genetic defect is obvious. However, in the rest of the cases and particularly where genetic screening was negative, the co-existence of a PA and a PHEO or PGL may indeed represent an extremely rare coincidence due to mutations in co-segregating genetic defects or epigenetic changes; few cases may be explained by ectopic hormone secretion by a PHEO/PGL mimicking a functioning PA. Regarding the reported variations in sporadic PAs in Table 3, most of them are variants of unknown significance and some of them are predicted to be damaging using the available prediction tools. However, one should perform functional studies to prove their deleterious effect.

Table 3.

SDHx variants/mutations in pituitary adenomas.

Patient Nr	Gene	Sex	Age	Type	Size/Treatment	ICH/LOH	Other tumors/conditions	Family history	Genetics tested/Prediction	Reference
1	SDHB	F	49	PRL	Micro/NK	SDHB positive	PHPT	Aunt: ACTH PA	c.5 C > T; p.A2V [Deleterious (SIFT)/benign (PolyPhen)]	de Sousa et al. (2017) [27]
2	SDHC	F	34	PRL	Micro/NK	SDHB positive	None	Brother: PRL PA	c.403 G > C; p.E110Q (Deleterious (SIFT)/possibly damaging (PolyPhen))	de Sousa et al. (2017) [27]
3	SDHC	F	63	GH	NK	SDHB positive	Pituitary gangliocytoma, PHPT	No	p.E110Q [Deleterious (SIFT)/possibly damaging (PolyPhen)]	de Sousa et al. (2017) [27]
4	SDHA	M	62	NFPA	Macro/TSS	Loss of staining for both SDHA and SDHB	None	No	c.725_736del/c.989_990insTA (double hit)	Gill et al. (2014) [18]
5	SDHB	F	35	PRL	Macro/TSS	intracytoplasmic vacuoles	None	Mother: Prolactinoma, brother: PGL	c.298 T > C (damaging mutation)	Dénes et al. (2015) [22]
6	SDHA	M	30	NFPA	Macro/TSS	ICH negative for SDHA and SDHB	None	Mother carotid body paraganglioma	c.1873C > T p.His625Tyr	Dwight et al. (2013) [24]
7	SDHB	F	15	Nl<	NK	NK	NK	NK	ex. 7 c.761insC p.254fsX255	Benn et al. (2006) [121]
8	SDHD	M	12	Cushing	Micro/TSS	Not studied	Thyroid	No	ex. 2 c.53 C > T, p.Ala18Val (very rare SNP (allele frequency <0.001) Benign (PSIC 0.004) (by PolyPhen) Deleterious (Score 0.01) (by SHIFT)	Xekouki et al. (2015) [26]
9	SDHB	F	31	PRL	Macro/TSS	LOH at SDHB locus in the PA/SDHB staining: loss of expression of SDHB	None	Grandmother’s first cousin PGL	del ex. 6 to 8	Dénes et al. (2015) [22]
10	SDHD	F	16	Cushing	Micro/TSS	Not studied	None	No	ex. 2c.149 A > G/p.His50Arg (Probably damaging (PSIC 0.993) (Polyphen) Benign (Score 0.48) (SHIFT)	Xekouki et al. (2015) [26]
11	SDHB	F	14	Cushing	Micro/TSS	Not studied	None	No	ex. 5c.487 T > C, p.Ser163Pro (reported as potentially pathogenic by Ni et al. 2012) [122]	Xekouki et al. (2015) [26]
12	SDHB	M	10	Cushing	Micro/TSS	Not studied	Brachydactyly/dysmorphic features	No	ex. 5c.487 T > C, p.Ser163Pro (reported as potentially pathogenic by Ni et al. 2012) [122]	Xekouki et al. (2015) [26]

M: Male; F: Female; NK: Not known; PRL: Prolactinoma; GH: Acromegaly; NFPA: Non-functional pituitary adenoma; Macro: Macroadenoma; Micro: Microadenoma; PA: Pituitary adenoma; PGL: Paraganglioma; PHPT: Primary hyperparathyroidism, TSS: Transphenoidal surgery.

Phenotypic and pathological characteristics of PAs with SDHx mutations

We attempted to see whether PAs with SDHx mutations appear to have a different progression as previously described [21, 22, 25, 26]. Unfortunately, not all clinical data were available to make a safe distinction between the PAs as part of 3PAs with and without germline mutations. However, based on the available data from the reported cases we looked for any difference in phenotypic characteristics between PAs in the context of 3PAs with (Table 1) and without genetic mutations (Table 2) and the isolated PAs with SDHx mutations/variants (Table 3). The SDHx-related PAs in 3PAs were more common among familial cases, more frequently macroadenomas, they often led to multiple phenotypes within the same family (somatotropinomas, prolactinomas, and nonfunctioning adenomas) and required more than one modes of treatment (Table 1 & Supl. Table 1S). Patients with isolated PAs and SDHx mutations/variants (Table 3) were significantly younger compared to the ones in cases with 3PAs regardless of the presence of a genetic mutation (Table 1, 2), and they all required surgery (regardless of the size of their tumors) but did not require multiple treatments (Supl. Table 4). A sub-analysis between the isolated PAs with SDHx mutations/variants and their counterparts in 3PAs, revealed that the latter were more frequently macroadenomas and required more than one treatment modality, which may suggest that the presence of PHEOs/PGLs may have contributed to the increased size and treatment resistance (Supl. Table 2S).

At this point, there does not appear to exist an apparent phenotype-genotype correlation, although very few families with 3PAS have other tumors associated with SDHx mutations such as gastrointestinal stromal tumors (GISTs) or renal cancer. The limited number of cases and the lack of prospective studies also do not allow for an accurate estimate of the expected age of presentation: based on the available cases (Table 1, 3) age at diagnosis ranged from as early as 15 years to as late as 72 years old. PHEOs or PGLs in these patients were bilateral and/or multiple, with a tendency to recur. The latter observation is one of the characteristics of SDHx-related PHEOs/PGLs [35].

Interestingly, in our first report, the growth hormone (GH) receptor (GHR) gene was found to be expressed in PHEO samples from our patient with the SDHD mutation, as well as in tumor samples from other patients harboring SDHB or SDHD mutations [15]. Based on this finding together with the clinical observation that there was a noticeable decrease (almost three-fold) of plasma and urinary metanephrines after pituitary transsphenoidal surgery (that was greater than the one noted following bilateral adrenalectomy for the patient’s PHEOs), we assumed that normalization of GH levels after TSS contributed significantly to such biochemical changes. This was the first and only study so far reporting the expression of GHR in PHEOs whereas there are reports of the differential expression of ghrelin and GH-releasing hormone (GHRH) receptors in various adrenal tumors, including PHEOs [36–38]. The role of GHR in SDHx-mutant tumors needs to be investigated further.

One interesting histological phenotype reported by Dénes et al. [22], in pituitary gland tissues from patients with SDHx mutations was the presence of intracytoplasmic vacuoles in their PAs. Although electron microscopy was not used to identify the exact nature of the vacuoles [39], there is a possibility that they represented autophagic bodies. The relationship between hypoxia-related pathways and autophagy activation is well established [40, 41] and autophagy has been shown to contribute to chemo- and radio-therapy resistance [42, 43]. We have described a similar morphological finding in the PA tissue of the original case with the SDHD mutation (Fig. 1a) [15].

Fig. 1.

a: H & E staining of the pituitary adenoma of the SDHD patient reported by Xekouki et al., 2012 [15, 17] showing cytoplasmic vacuoles (arrow) (× 40) b: Enlarged mitochondria with marked swelling and abnormal and/or missing internal cristae in pituitary cells of Sdhb + / − mice (electron microscopy 200 nm).

Loss of SDHB immunochemistry has been shown to be an excellent indicator of germline or somatic mutations in the SDHx genes [1, 44, 45]. In most of the SDHx-related PAs reported so far, completely absent or weak diffuse staining of SDHB was found (in those cases that this was performed) (Table 1, 3). Therefore, as in PHEOs/PGLs, SDHB staining may be used as an additional diagnostic tool to screen for SDHx mutations. Furthermore, as shown by Richter et al. [46], profiling of Krebs cycle metabolites, such as ratios of succinate:fumarate with the use of mass spectrometry, is another useful method to identify patients for testing of SDHx mutations and to assess functionality associated with SDHx variants of uncertain significance in PGLs. Although this method has not been used in any of the 3PAs cases it would be interesting to see whether it could predict the presence of SDHx mutations or distinguish damaging mutations from nonfunctional polymorphisms in pituitary adenomas in the context of 3PAs as in PGLs.

Traditionally LOH has been used in oncology to confirm the causative association between a tumor and the loss of a tumor suppressor gene [47]. As shown in Table 1, in some cases no LOH studies were performed; in few cases no LOH of the SDHx mutated locus was identified. Does this mean that the absence of consistent LOH in PAs indicates lack of association with the identified SDHx mutations? Although this is difficult to answer with certainty at this time, bilateral adrenal medullary hyperplasia associated with a germline SDHB mutation showed retention of heterozygosity [48], and PHEOs without loss of the normal SDHD allele have been shown in patients with pathogenic SDHD mutations [49]. Additionally, cases of “paradoxical” loss of the mutant SDHx allele have been shown [24, 50], pointing to the suggestion that the SDHx defects may not always lead to tumorigenesis in the classical tumor suppressor gene way. Finally, epigenetic alterations such as somatic SDHC promoter methylation and postzygotic somatic mosaicism could provide another explanation for those cases negative for germline mutations [51, 52].

Molecular mechanisms of SDHx mutations in the development of a PA

To further understand the causal relationship between PA and SDHx mutations, we analyzed the pituitaries of adult Sdhb^{+ / −} mice. Although no gross pituitary tumors were detected, the pituitary of 12-month old Sdhb ^{+ /−} mice was hypercellular, mainly due to the increased number of prolactin (PRL)-secreting cells and to a lesser extend GH-secreting cells. There were also blood-filled lakes, abnormal mitochondria (Fig. 1b), nuclear inclusions of unknown nature, and abnormal and poorly defined heterochromatin forming a ring-shaped pattern rather than being centrally located as in pituitaries of the normal control animals [26]. Pituitary hyperplasia likely preceding the formation of adenomas, was found in 3- and 12-month-old Aip-deficient male mice [53] as well as in those lacking p19 ^(arf) or p27^(kip–1), and those overexpressing the beta-catenin or the pituitary tumor-transforming gene (CTNNB1 and PTTG, respectively) [54–57]. In humans, pituitary hyperplasia has been reported in patients with germline mutations in the PRKAR1A or AIP tumor suppressor genes, and in those with gigantism and chromo-some Xq26 duplications [58–60].

Like in other SDHx-deficient PHEOs/PGLs [61], there is evidence that hypoxic signalling is activated in SDHx-mutated pituitary tumors [15]. This finding was supported by the HIF-1α strong cytoplasmic (and occasionally nuclear) expression in the pituitaries of Sdhb + /− mice [26]. Other evidence of hypoxia-activated pathway in these mice was the enlarged mitochondria with destructed cristae found in the Sdhb + /− animals [26, 62]. Similar mitochondrial findings have been reported in tumor samples from patients with Carney triad, as well as in paragangliomas from patients with SDHC and SDHD mutations [56]. Fragmented and defective mitochondria have been found in many different cancers and alterations in mitochondrial dynamics are associated with tumor progression or resistance to therapy [63]. The altered chromatin pattern that we also observed in Sdhb + /− mice, like the one observed in mice overexpressing PTTG and in hypoxic cells [55, 64] is another indication that chronic activation of pseudohypoxia in SDHx-deficient pituitary cells can drive genetic instability and eventually lead to tumor formation.

Although electron microscopy failed to identify the actual nature of the nuclear inclusions seen in Sdhb ^{+ /−} pituitary cells with light microscopy, it has revealed that these appeared to be fused to the nucleus rather than being intranuclear. These inclusions resembled the vacuoles described by Denes et al., in SDHx-deficient pituitary tumors and may represent late autophagic vacuoles previously filled with digested abnormal mitochondria. Thus, we may speculate that PA formation from SDHx mutations may follow a long process, whereby hyperplasia is the initial response to a prolonged stimulation due to activation of pseudo-hypoxia signals (Fig. 2).

Fig. 2.

Proposed mechanism of tumor development in SDHx-deficient pituitary adenomas; SDHx mutations may initially lead to hyperplasia, as it was shown in Sdhb + / − mice, which eventually either at the presence of another mutation or epigenetic events it will lead to pituitary tumor development (H & E staining of the original case reported in 2012).

Recommendations for genetic screening and follow-up of patients with 3PAs

Although cases of 3PAs have been reported since 1952, it was only in 2012 when we showed that SDHx mutations may be involved in pituitary tumorigenesis [15]. So far there are 28 cases of confirmed SDHx mutation-related pituitary adenomas (with the variants of unknown significance excluded). Currently, we recommend that a detailed baseline medical and family history is taken from all these patients along with careful physical examination to detect signs of other tumors associated with SDHx defects (for examplle, GISTs) (Fig. 3). Hormonal testing should be obtained, once familial PHEOs and/or PGLs are recognized in the index case and their family members. Attention should be paid to any symptoms that would indicate GH or PRL hypersecretion and visual disturbance, as most of the pituitary adenomas in the context of 3PAs are PRL- or GH-secreting macro-adenomas or NFPAs. If there are no suspicious symptoms or findings and biochemistry is normal then annual biochemical surveillance should include testing for PHEOs/PGLs, as per the most recent recommendations [65, 66]. If at any time, clinical findings or biochemistry indicate hormonal hypersecretion, a pituitary MRI should be performed.

Fig. 3.

Recommended clinical investigation and follow-up in patients with 3PAs.

Treatment of PAs due to SDHx mutations should not differ from the recommended treatment for sporadic tumors [67–71]. However, it is possible that more than one treatment modalities will be required, as these tumors are often more aggressive and tend to recur more frequently than their sporadic counterparts (although this impression may change as more cases are now identified by screening and are followed prospectively).

Special attention should be given to patients that may have MEN1, MEN2, or MEN4, as now we know that all three syndromes may be associated with both PAs and PHEOs/PGLs. Thus, plasma or urine metanephrine measurements may be needed prior to surgeries in patients with suggestive signs and symptoms of a PHEO or a PGL. Finally, in case of a family history of a PHEO/PGL due to MAX mutation, careful medical history and physical examination should be performed in all mutation carriers for the presence of other tumors related to MAX mutations such as PA and renal cancer as recently described [33, 72].

Genetic testing for SDHx mutations should be performed in any patient who presents with 3PAs particularly if there is family history of PHEOs/PGLs in other family members (the latter may not only include first-degree relatives). In the absence of a family history of PHEOs/or PGLs, screening for SDHx mutations in patients with apparently sporadic 3PAS may also be performed, as these genetic defects are known to have low penetrance in affected families. Alternatively, and/or if genetic screening is not readily available, SDHB staining and succinate:fumarate ratio maybe another way to look for SDHx defects. The presence of intracytoplasmic vacuoles is another pathology finding strongly suggestive of SDHx gene defect. If screening for SDHx mutations is negative or pathology and biochemistry are not suggestive of SDHx defects, then screening for menin or RET mutations should be based on the presence of clinical features and/or tumors that would indicate the presence of the MEN1 or MEN2 syndromes. Screening for MEN4 is not considered routine at this point, as is the case for other genes known to predispose to PA development [73]. Finally, screening for MAX mutations should be considered if SDHx and menin are screened negative.

Final remarks

In the era of next generation, sequencing and new molecular techniques hundreds of new genes responsible for tumors have been discovered [74]. The opposite is also true: many new associations are found for mutations of the same gene(s). Indeed, today is the era of physician-researchers working like crime scene investigation (CSI) agents, gathering data carefully, connecting seemingly unrelated facts, and discovering new phenotypes for known gene defects [75]. The SDHx genes are very well suited for the investigation of new associations, as they are expressed in all cells, have an essential role in mitochondrial function, and their defects have low, overall, penetrance.