Clinicopathological and genomic analysis of pediatric pheochromocytoma and sympathetic paraganglioma

Abstract

Pediatric patients with pheochromocytoma (PCC)/paraganglioma (PGL) (PPGL) are rare, and clinicopathological investigations, especially the relationship between gene analysis and histological features, are insufficient. We comprehensively examined the clinical data, germline/somatic variants (mutations), and pathological characteristics of operated tumors using immunohistochemical expression and histological grading by Grading of Adrenal PCC and PGL score. This study included 28 patients (15 males and 13 females) aged <19 years. The age at the diagnosis was 12.8 ± 4.5 years. The included patient often had multiple PPGLs, with 39 tumors, including 21 PCCs and 18 PGLs, with average tumor sizes of 45.0 ± 22.8 and 42.6 ± 23.6 mm, respectively. Genomic types examined by gene mutations and immunohistochemistry of CA9 for VHL, SDHB for SDHx, and MAX for MAX, classified them into 14 VHL (50%), ten SDHx (35.7%), one MAX (3.6%), and three unknown (10.7%) types. Tumor metastasis was limited to two SDHB-related PPGLs, but not to VHL-related PPGLs. In both patients, the metastatic sites were the bones. The average GAPP score of the PPGLs was 2.9 ± 1.5 in VHL and 5.3 ± 1.7 in SDHB, and histological grades were well-differentiated in VHL and moderately differentiated in SDHB. SSTR2 expression was positive in 90% of SDHB-related PPGLs, but negative in 75% and weakly or focally positive in 25% of VHL-related PPGLs. Most pediatric PPGLs (90%) demonstrated mutations in VHL, SDHB, and MAX, with histological features depending on the mutation type. Combined genetic and immunohistochemical examination is desirable for accurate genomic diagnosis, and clinicopathological study.

Graphical Abstract

Introduction

Pheochromocytoma (PCC) and sympathetic paraganglioma (PGL) are catecholamine-producing non-epithelial neuroendocrine tumors derived from the neural crest. PCC, which is an adrenal sympathetic PGL and an extra-adrenal sympathetic PGL are called PPGL altogether.

The prevalence of PPGLs is estimated between 1:6500 and 1:2500 (15–40/100,000), with an annual incidence of 500 to 1,600 cases per year in the United States [1]. The incidence of pediatric PPGLs is approximately 20% of all cases [2, 3] and pediatric PPGLs are especially related to familial manifestations with genetic susceptibility syndromes, such as von Hippel-Lindau syndrome (VHL) and hereditary pheochromocytoma-paraganglioma syndrome (HPPS) [3, 4]. Although approximately 10% of patients have a positive family history, genetic screening has revealed that approximately 30% have germline variants (mutations) in any of the identified genes [5]. The reported inheritance changed from 30% to 40% in small pediatric case series to 80% in a larger series [6, 7]. In addition, recent genetic examinations revealed that somatic mutations in hereditary susceptibility genes are found in as many as 20% of truly sporadic PPGL cases without indication of heritable disease [8, 9].

Although clinical and genetic investigations of pediatric PPGLs have been extensively reviewed [2, 10-12], pathological studies of resected tumors correlated with genetic mutations in pediatric PPGLs are limited. Carboxyl anhydrase 9 (CA9), succinate dehydrogenase B (SDHB), and MYC-associated factor X (MAX) are established biomarkers for immunohistochemistry (IHC) of PPGLs related to VHL, SDHx (SDHA, ADHAF2, SDHB, SDHC, and SDHD), and MAX, respectively [13-17]. In addition, somatostatin analogs have been used for treating metastatic PPGLs [18]; therefore, we examined the expression of somatostatin receptor type 2 (SSTR2). In our collaborative study on PPGLs in Japan, we included 28 pediatric patients <19 years. We comprehensively examined the clinical data, genetic mutations, and histological characteristics of the operated tumors, including gross findings, such as multiple intra-adrenal tumors and/or metastatic tumors, and microscopic findings, such as immunohistochemical expression of CA9 for VHL, SDHB for SDHx, MAX for MAX gene mutations, SSTR2, and histological analysis for predicting prognosis using the Grading for Adrenal Pheochromocytoma and Paraganglioma (GAPP) score [19]. Here, we propose the characteristics of pediatric PPGLs from the aspects of genetic types, histopathology, possibility of metastasis, and treatment with somatostatin analogs.

Materials and Methods

Patient clinical characteristics and tumor types

Our collaborative study group for pediatric PPGLs in Japan included 28 pediatric patients younger than 19 years from 2011 to 2023. The youngest age at diagnosis of patients with tumors with several occurrences and/or recurrences, was described. The average age of the patients at diagnosis was 12.8 ± 4.5 (range: 1.3–19) years, with 15 males and 13 females. The chief complaints were headache, hypertension, and sweating in 40%, 25%, and 21% of the patients, respectively. Twenty-six patients had elevated catecholamine levels (functional type), and two patients had normal catecholamine levels (non-functional type). In functional type, 2 patients had elevated epinephrine (E) + norepinephrine (NE) + dopamine (DA), 2 patients had E + NE, 7 patients had NE + DA, and 15 patients had only elevated NE. Age-matched reference values for catecholamine were not examined. The total number of tumors was 39, including 21 PCCs with seven bilateral (6 VHL and 1 MAX) and 18 PGLs, with 13 located in the retroperitoneum (2 patients had 2 PGLs), 2 in the urinary bladder, and 1 in the mediastinum. The tumor size was 45.0 ± 22.8 (7–95) mm in PCC and 42.6 ± 23.6 (10–85) mm in PGL. Two patients had pelvic bone metastases: both patients had SDHB-related PCCs, one patient was found to have metastasis during the first operation for PCC, and the other patient was found to have metastasis 4 months after the first operation, as described in their clinical records.

Clinical data, such as sex, age at diagnosis, clinical complaints, elevated catecholamine types, number of tumors, tumor location, and tumor size, are summarized in Table 1.

Table 1. Clinical characteristics of 28 pediatric patients with pheochromocytoma and paragangliomas.

Characteristics		Overall (patients) (n = 28)	PCC (bilateral)	PGL (two)	PCC, bilateral + PGL
Gender	Male	15 (54%)	9	5	1
	Female	13 (46%)	3	8	2
Age at diagnosis (years)		12.8 ± 4.5, range (1.3–19)
Clinical complaints
	headache	11 (40%)	4	7
	hypertension	7 (25%)	4	3
	sweating	6 (21%)	4	2
	vomiting	3 (11%)	2	1
	no complaints	4 (14%)	2	2
Catecholamine types
	E + NE + DA	2 (7%)	2	0
	E + NE	2 (7%)	2	0
	NE + DA	7 (25%)	1	5	1
	NE only	15 (54%)	7	7	1
	Non-functional	2 (7%)	0	1	1
Number of tumors		39	21	18
Tumor location			Adrenal gland: 14 (7)	Retroperitoneum: 13 (2)
				Urinary bladder: 2
				Mediastinum: 1
Tumor size, mm			45.0 ± 22.8, range: 7–95	42.6 ± 23.6, range: 10–85

Abbreviations: PCC: pheochromocytoma, PGL: paraganglioma, E: epinephrine, NE: norepinephrine, DA: dopamine

Family history and genetic analysis

Family history confirmed that 8 patients had family history of PPGLs including 5 VHL, 1 with SDHB, 1 with MAX, and 1 with unknown gene type (Table 2).

Table 2. Genetic analysis and family history of pediatric PPGLs (Case numbers in Tables 2 and 3 are of the same patient).

Cases	Tumor type	Germline mutation	Somatic mutation	Family history
1	PCC	Negative	VHL c.514_522dupCCTGAGAAT, p.172P_174Ndup	Negative
2	PCC	Negative	VHL c.482G>A, p.R161Q	n.e.
3	PCC, bilateral & PGL	VHL c.482G>A, p.R161Q	Negative	Negative
4	PCC	VHL c.496 G>T, p.V166F + SDHD c.331 G>A, p.V111I	n.e.	Yes
5	PCC	Negative	VHL c.284C>G, p.P95R	n.e.
6	PCC	VHL c.499C>T, p.R167W	n.e.	Yes
7	PCC, bilateral	VHL c.371C>T, p.T124I	n.e.	Yes
8	PCC, bilateral	n.e.	n.e.	Yes
9	PGL, two	No mutation on SDHx, &VHL	n.e.	Negative
10	PCC, bilateral & PGL	n.e.	n.e.	n.e.
11	PCC, bilateral	n.e.	n.e.	n.e.
12	PGL	VHL c.250G>A, p.V84M	n.e.	n.e.
13	PGL	VHL c.129T>C, p.L386P	n.e.	Yes
14	PCC, bilateral	VHL c.293A>G, p.Y98C	n.e.	Yes
15	PCC	Negative	VHL c.595delG, p.E199fs	Negative
16	PGL	n.e.	n.e.	n.e.
17	PGL	n.e	n.e.	n.e
18	PGL	SDHB c.423 + 1G>A	SDHB LOH	Negative, de novo
19	PGL	SDHB c.424-2delA	SDHB LOH	Yes
20	PGL	n.e.	n.e.	n.e.
21	PCC	SDHB c.201-2A>C	n.e.	n.e.
22	PGL	SDHB c.201-2A>C	n.e.	n.e.
23	PGL	SDHB c.137G>A, p.R46Q	SDHB LOH	n.e.
24	PCC	n.e.	n.e.	n.e.
25	PCC	n.e.	n.e.	n.e.
26	PCC, bilateral & PGL	MAX c.223 C>T, p.R75*	n.e.	Yes
27	PGL,two	SDHB c.470delT, L157fs	n.e.	Negative
28	PGL	SDHB c.470delT, L157fs	n.e.	Negative

Abbreviations: PCC: Pheochromocytoma, PGL: Paraganglioma, n.e.: not examined

Genetic analysis was performed using Sanger sequencing for SDHB, SDHD, VHL, RET, and MAX, and/or multiplex ligation-dependent probe amplification (MLPA) analysis for SDHB, SDHC, SDHAF1, SDHAF2, and VHL, as previously described [20, 21]. Written informed consent was obtained from all patients.

Immunohistochemistry

Tumors were fixed in 10% buffered formalin and embedded in paraffin. All surgically resected PPGLs were initially diagnosed by pathologists in their own institutes; however, special issues such as GAPP and IHC for tyrosine hydroxylase (TH), dopamine beta-hydroxylase (DBH), SDHB, CA9, MAX, and SSTR were not objectives for general pathologists. Subsequently, all clinicopathological issues of the above were extensively examined in National Hospital Organization Hakodate Medical Center. After reviewing all the sections stained with hematoxylin and eosin, sections suitable for IHC were selected. IHC for chromogranin A (CgA), synaptophysin, CD56, S100 protein, TH, DBH, and Ki67 were performed on all tumors to confirm that they were catecholamine-producing, or non-producing PPGLs. In addition, we used antibodies to characterize genomic mutations in PPGLs, such as CA9 for VHL [13, 14], SDHB for SDHx [15, 16], and MAX for MAX [17]. The expression of somatostatin receptor type 2 (SSTR2) was examined using IHC as a somatostatin analog for the treatment of metastatic PPGLs [18]. IHC was performed using a VENTANA BenchMark ULTRA Slide Staining System (Roche Diagnostics, Indianapolis, IN, USA) with primary antibodies against CgA (DAKO; DAK-A3, 1:100), synaptophysin (Nichirei, Tokyo, Japan; clone 27G12, prediluted), CD56 (Nichirei clone MRQ42, prediluted), TH (Chemicon, Temecula, CA, USA; rabbit polyclonal, 1:500), DBH (Abcam, clone EPR209487, 1:3000), Ki67 (MIB1) (Ventana; clone 30-9, prediluted), S100 (Nichirei, rabbit polyclonal, prediluted), CD31 (DAKO, clone JC70, prediluted ), alpha smooth muscle actin (DAKO, clone 1A4, prediluted), CA9 (Abcam, clone15086, 1:100), SDHB (Sigma, clone HPA002868, 1:200), MAX (Santacruz, SC-197, 1:600), and SSTR2 (Nichirei, clone EP149, prediluted). Appropriate positive controls, such as the human adrenal medulla for CgA, synaptophysin, CD56, S100, TH, and DBH [22] pancreatic tissue [23, 24] for SSTR2, and duodenal mucosa [25] for CA9, were added. IHC-positive locations were on the cell membrane for CA9 [13, 14] and SSTR2 [26], in the cytoplasm for SDHB [15, 16] and in the nucleus for MAX [27]. PBS was used as a negative control. Tissue localization for identifying VHL, SDHB, and MAX mutations using IHC was as follows: CA9-IHC positivity localized on the cell membrane [13, 14], SDHB-IHC negativity in the cytoplasm [15, 16] and MAX-IHC [27] negativity in the nucleus provided supportive evidence for mutations in VHL, SDHB, and MAX, respectively.

Definitions for immunohistochemical grading of CA9 and SSTR2 were as follows: 0, negative; no positive cells; 1+, less than 10% positive cells; 2+, less than 50% positive cells; 3+, more than 50% positive cells; SDHB, 0, negative; 1+, negative or very weakly positive; 2+, focally positive; and 3+, diffusely positive.

Genomic types were defined by the presence of gene mutations and/or immunohistochemistry.

Statistical analyses were performed using StatMate V Software (Takahashi Y, ATMS, Tokyo, Japan).

The histological analysis and scoring for histological grading were performed based on the previously published reference of GAPP [19].

Results

Genomic types by gene analysis and immunohistochemistry

Genetic analysis revealed that 11 patients had VHL-related mutations, of which seven had germline mutations and four had somatic mutations without germline mutations (Table 2). CA9-IHC positivity localized to the cell membrane provides supportive evidence for mutations in VHL [13, 14]. CA9-IHC positivity was localized on the cell membrane of six of the nine VHL-related PPGLs and three cases without gene analysis (cases 8, 11, and 10). Among these cases, cases 8 and 11 had a family history of PCC, and case 10 had bilateral PCCs with urinary PGL, all of which were compatible with hereditary PPGLs. Therefore, these three PPGLs were confirmed as VHL-related PPGLs. In contrast, three VHL-related PPGLs (cases 4, 5, and 12) confirmed by gene analysis showed no immunoreactivity to CA9. Overall, VHL-related PPGLs were 14 cases (50%), including 6 cases positive for both VHL-gene mutations and CA9-IHC, three positive VHL-gene mutations but negative for CA9-IHC, three negative VHL-gene mutations but positive for CA9-IHC, and two positive VHL-gene mutations without CA9-IHC. There was an exceptional case (case 4), which had combined mutations in VHL and SDHD and a family history of the same gene mutations. However, the patient had no head and neck PGLs, which are usually associated with SDHD, or positive cytoplasmic staining for SDHB; therefore, we included the patient in the VHL-group. The 14 VHL-related PPGLs included 10 PCCs, 2 PGLs, and 2 PCC + PGL.

Genetic analysis revealed that 7 patients had SDHB-related germline mutations.

SDHB-IHC negativity is established evidence for SDHB-related PPGLs [15, 16]. SDHB-IHC-negative staining was observed in all seven SDHB-related PPGLs. In addition, 3 cases of unexamined gene mutations were negative for SDHB-IHC and compatible with SDHB-related PPGLs. In total, SDHB-related PPGLs were identified in 10 cases (35.7%), including 8 PGLs and 2 PCCs. In addition, 4 VHL-related PPGLs (two germline and two somatic mutations) were negative for SDHB-IHC, and these 4 PPGLs were previously suspected to be SDHB-related and recommended to undergo SDHx-gene analysis.

Genetic analysis revealed that one patient had MAX-related germline mutations.

Loss of nuclear staining by MAX-IHC provided supportive evidence for MAX-related PPGLs [17]. In one MAX-related PCC, tumor cell nuclei were negative using MAX-IHC, in contrast to the positive staining of the adjacent adrenal cortical cells. Both gene mutations and MAX-IHC confirmed MAX-related PCC.

Finally, the combined gene analysis and immunohistochemistry resulted in 14 VHL (50%), 10 SDHB (35.7%), 1 MAX (3.6%), and 3 mutation-unknown PPGLs (10.7%) in 28 pediatric patients (Table 3).

Table 3. Immunohistochemistry, GAPP score, and genomic type of pediatric PPGLs (Case numbers in Tables 2 and 3 are of the same patient).

Cases	CA-9 IHC	SDHB-IHC	SSTR2-IHC	GAPP score	Genomic type
1	n.e.	n.e.	n.e.	5	VHL
2	1+	Negative	n.e.	2	VHL
3	3+	2+	0	3	VHL
4	Negative	3+	n.e.	2	VHL
5	Negative	3+	0	1	VHL
6	3+	2+	0	4	VHL
7	2+	Negative	1+	5	VHL
8	3+	2+	0	1	VHL by IHC
9	Negative	2+	0	3	Not determined
10	1+	3+	n.e.	2	VHL by IHC
11	2+	3+	n.e.	0	VHL by IHC
12	Negative	2+	1+	3	VHL
13	2+	Negative	0	5	VHL
14	2+	3+	n.e.	4	VHL
15	n.e.	Negative	n.e.	5	VHL
16	Negative	3+	3+	5	Not determined
17	Negative	Negative	2+	4	SDHB by IHC
18	Negative	Negative	3+	9	SDHB
19	n.e.	Negative	n.e.	6	SDHB
20	Negative	Negative	2+	5	SDHB by IHC
21	Negative	Negative	2+	4	SDHB
22	Negative	Negative	3+	5	SDHB
23	Negative	Negative	0	4	SDHB
24	Negative	Negative	n.e.	3	SDHB by IHC
25	Negative	3+	0		Not determined
26	Negative	3+	3+	3	MAX
27	Negative	Negative	3+	6	SDHB
28	Negative	Negative	3+	6	SDHB

Abbreviations: PCC: Pheochromocytoma, PGL: Paraganglioma, n.e.: Not examined

Somatostatin receptor Type 2 immunohistochemistry

Immunohistochemically, SSTR2 is localized to the circumferential membrane of tumor cells [26]. SSTR2-IHC was negative in 6 (75%) and very focally positive in 2 (25%) of 8 VHL-related PPGLs. However, 8 out of 9 (89%) SDHB-related PPGLs examined were positive for SSTR2-IHC. Therefore, SSTR2-IHC was almost negative in VHL-related PPGLs, in contrast to the strongly positive SDHB-related PPGLs.

Combined IHC of CA9, SDHB, and SSTR2 revealed characteristic features: VHL-related PPGLs were positive for CA9 and SDHB, but negative or weakly positive for SSTR2. In contrast, SDHB-related PGL was negative for SDHB and CA9, but strongly positive for SSTR2. MAX-related PCC was negative for MAX-IHC and CA9-IHC and positive for SDHB-IHC and SSTR2-IHC. Therefore, the combined IHC of CA9, SDHB, MAX, and SSTR2 was useful for screening genomic mutations in VHL, SDHB, and MAX in pediatric PPGLs.

Histological grading using the GAPP score

The histology of the PPGLs was analyzed using the GAPP score. The GAPP score is composed of six parameters, including histological pattern, cellularity, comedo necrosis, vascular or capsular invasion, Ki67 labelling index, and catecholamine type, and the histological grade is subsequently classified into three types: well, moderately, and poorly differentiated, depending on the total score of the parameters [19]. GAPP grading correlates with the patient prognosis of adult PPGLs and was validated [28, 29]; however, pediatric PPGLs have not been evaluated using GAPP grading. In this study, we compared the statistical significance of GAPP scores in VHL- and SDHB-related PPGLs using the ztTEST. The number of tumors in VHL- and SDHB-related PPGLs was 14 and ten, respectively. The GAPP scores in PPGLs were 2.9 ± 1.5 in VHL and 5.3 ± 1.7 in SDHB (p < 0.01, p = 0.00154). The GAPP score was defined as 0–2 points for the well-differentiated type, 3–6 for the moderate differentiated type, and 7–10 for the poorly differentiated type. Thus, the histological grade was well-differentiated in VHL and moderately differentiated in SDHB. None of the pediatric PPGLs were poorly differentiated. Some cases in this study are previously reported case reports [30-32]. The results of IHC for CA9, SDHB, MAX, SSTR2, and GAPP score are summarized in Table 3.

Histological characteristics of VHL, SDHB, and MAX-related PPGLs

VHL-related PPGLs had multiple nodules in gross, even in a single adrenal gland. The histology of VHL-related PPGLs usually demonstrated a regular Zellballen pattern composed of several clusters of chief cells positive for CgA, TH, and DBH. The Zellballen pattern was enhanced by well-developed S100-positive sustentacular cells and surrounding capillaries with swelling endothelial nuclei. In some cases, areas of larger Zellballen patterns were mixed with areas of small Zellballen patterns, particularly in larger tumors. CA9-IHC has been reported to be specifically positive on cell membrane of VHL-related PPGLs; however, CA9-immunoreactivity showed various numbers of positive cells and intensity for each case, and negative in some cases. The pseudo-rosette formation pattern was occasionally observed in VHL-related PPGLs (Fig. 1).

Fig. 1. VHL-related PPGL.

A. Tumor is composed of small irregular-shaped cell nests associated with well-developed vascular vessels.

B. Tumor cells are positive for chromogranin A, associated with well-developed vascular vessels.

C. Tumor cell nest, which is surrounded by well-developed S100 protein-positive sustentacular cells, is named Zellballen pattern.

D. Tumor nodularity is enhanced by vessels stained with aSMA-IHC.

E. CA9 is a specific marker for VHL-PPGLs. CA9 reacts on cell membrane. This case shows very few positive CA9 cells.

F. CA9 positive cells are more evident in this tumor.

G. CA9-IHC strongly demonstrates a positive reaction in this tumor.

H. SSTR2-IHC is negative or focally positive in this type of PPGL.

I. A pseudo-rosette pattern is occasionally observed in VHL-related PPGLs.

SDHB-related PPGLs were usually single nodules in the adrenal gland or mostly in the extra-adrenal paraganglia. SDHB-related PPGLs usually have small and large irregular Zellballen patterns. Tumor cells of this type were characteristically negative for SDHB-IHC, and positive for SSTR2. A pseudo-rosette pattern, which showed an arrangement of monotonous cuboidal tumor cells around vessels, also known as a pseudopapillary pattern, was sometimes observed in SDHB-related PPGLs [33]. Although CgA or DBH-IHC is usually stained diffusely in the cytoplasm; however, in cases with a pseudo-rosette pattern, CgA or DBH-IHC characteristically demonstrated subnuclear location, that is, pseudo-rosette-forming PPGLs had polarity of cell arrangement which is a common feature of epithelial neuroendocrine tumors, such as carcinoid tumors. This pattern has been reported as one of the characteristic features of SDHB-related PPGLs [34]; however, a pseudo-rosette pattern was occasionally observed in other germline mutations, such as VHL, as shown in Fig. 1, and RET [35]. Pseudo-rosette formation may be a feature of tumors derived from some genetic mutations, especially SDHB (Fig. 2).

Fig. 2. SDHB-related PGL.

A. Tumor shows small and large irregular Zellballen patterns, which are composed of monotonous proliferation of tumor cells divided by irregular proliferation of vascular vessels.

B. Tumor cells are positive for chromogranin A in the left half area; however, the right half area is negative for CgA. These findings are explained by tumor cell polarity, as CgA locates subnuclear lesions and positive areas show the subnuclear areas that have CgA granules; however, negative areas show the supranuclear area that has no CgA granules. This phenomenon is often observed in epithelial neuroendocrine tumors such as carcinoids.

C. This type of PPGL shows very few S100 positive sustentacular cells.

D. SDHB-related PPGL is characterized by SDHB-IHC negativity, although vascular endothelial cells are positive if they are not strongly reactive.

E. Somatostatin receptor type 2 is usually strongly positive.

F. The pseudo-rosette pattern of PPGL is characterized by cell arrangement around vessels.

G. In the pseudorosette-forming type, both dopamine-beta hydroxylase (DBH)-IHC and CgA-IHC demonstrate a subnuclear location of immunoreactivity, which is usually observed in carcinoid tumors.

MAX-related PCC were bilateral PCCs with a dumbbell type appearance in the adrenal gland with diffuse or nodular hyperplasia of the adrenal medulla [27]. Histology of MAX-related PCC showed diffuse proliferation of tumor cells of the basophilic cytoplasm with a high nuclear/cytoplasmic ratio. Tumor cell nuclei were negative for MAX-IHC, in contrast to positive nuclei of infiltrating lymphocytes. SSTR2-IHC were positive on tumor cell membrane, and SDHB-IHC demonstrated cytoplasmic rough granules compatible with mitochondria (Fig. 3).

Fig. 3. MAX-related PCC.

A. Bilateral PCC in MAX-related PCCs. Three isolated tumors in the right adrenal gland, and one tumor in the left adrenal gland.

B. Tumor cells with basophilic cytoplasm diffusely proliferate.

C. Tumor cell nuclei are negative for MAX-IHC in contrast to positive staining in infiltrating lymphocytes as an internal positive control.

D. Somatostatin receptor type 2 reacted on the cytoplasmic membrane of tumor cells.

E. SDHB-IHC shows granular cytoplasmic staining demonstrating mitochondriae, which ruled out SDHB-related PCC.

Metastasis

In this study, tumor metastasis was observed in 2 patients with SDHB-related PPGLs, but not in those with VHL-related PPGLs. Among the 2 patients with metastasis, 1 had PCC and the other had PGL. Both patients had bone metastases: one had multiple bone metastases at the time of the first operation, and the other had iliac bone metastasis four months after the operation for PCC. Both PPGLs were negative on SDHB-IHC and positive on SSTR2 staining.

Discussion

In the current study, we compared gene analysis and IHC expressions of CA9, SDHB, and MAX to determine genomic types, and then we examined correlation between genomic types and histological characteristics of VHL-, SDHB-, and MAX-related PPGLs, GAPP scoring, and SSTR2 expression (Graphical Abstract).

Graphical Abstract.

Gene analysis and IHC

Genetic testing of PPGLs associated with IHC characterization is an ideal method for identifying patients based on their clinical presentation and catecholamine secretory phenotypes. Considering the cost and convenience of genetic diagnosis in PPGL, it is evident that IHC can be useful in guiding the sequencing of SDH, MAX, VHL, and fumarate hydratase (FA) genes. IHC has been proposed as a guide for the biological validation of germline and somatic variants [36, 37]. Although IHC is a convenient tool for identifying the genetic backgrounds of susceptibility genes for PPGLs, the procedures of IHC should be carefully established to ensure suitable fixation of tumor tissues, a minimal but sufficient fixation time in neutral buffered formaldehyde, minimal time after cutting of paraffin-embedded tissues, and suitable positive controls for IHC, as basic issues for the IHC technique. For example, CA9-IHC for VHL-related PPGLs showed a vast variety of immunostaining (heterogeneity), and only one positive cell should be considered evidence of VHL even when most cells are negative [36], which is quite difficult and may be a cause of sampling errors. In this study, 3 VHL cases were negative for CA9-IHC; thus, several other sections were stained to avoid sampling errors. For example, SDHB-IHC sometimes demonstrates negative or very weak reactivity, causing disagreements in diagnosis [28]. In the present study, four VHL-related PCCs were negative for SDHB-IHC. The weak/negative SDHB-IHC results in VHL-PPGLs can be explained using transcription data as follows: (a) tumors with mutations in VHL, SDHB, and SDHD share a transcription signature of hypoxia, angiogenesis, and oxidoreductase imbalance; (b) SDHB protein is suppressed in tumors with mutations in SDHB and SDHC and in a subset of tumors with VHL mutations; and (c) HIF1a is involved in the SDHB downregulation observed in these tumors [38]. Thus, SDHB-IHC negative cases should be confirmed by genetic analysis to determine whether they are SDH or VHL-related PPGLs. Furthermore, SDHB-IHC negative cases may detect PPGLs with large scale deletions of the SDHB gene, representing approximately 10% of SDH-PPGLs. In such cases, there is inconsistency in SDHB-IHC and SDHB gene mutations analyzed using direct sequencing, and two methods for detecting large genomic mutations or duplications, including multiplex ligation-dependent probe amplification (MLPA), and quantitative multiplex PCR of short fluorescent fragments, can be performed to validate these findings [39]. In the case of undetermined vague IHC staining, gene analysis should be performed. Thus, combined use of gene analysis and IHC is ideal for precise diagnosis.

Histology and gene mutations

There are a limited number of previous studies on the histological characteristics of PPGLs. Our histological data on pediatric PPGLs associated with VHL and SDHB revealed different histological grades of GAPP, such as well-differentiated vs. moderately differentiated types, indicating different prognoses [19]. For the parameters in GAPP scoring, there are two types of Zellballen pattern; one is regular which is composed of uniform nests of tumor cells, and the other is composed of irregular Zellballen pattern indicated a mixture of small and large irregular tumor cell nests in which the size of the larger nests was at least ten times that of the smaller nests [19]. This difference in the Zellballen pattern was explained by micro vessel density labeled by CD31/CD34 [40-42] or αSMA [41]. Favier et al. [41] noticed that there were two different patterns, mainly benign PCCs having more regular patterns, with short and straight vascular segments distributed regularly over large areas of tumor tissue. The vascular density of these tumors was equivalent to that observed in normal adrenal medulla, while vascular endothelial cells of PPGLs have nuclei with more swelling than those of normal adrenal medulla. Gao et al. [43] classified PPGLs into low grade (well-differentiated type) and moderate grade (moderately differentiated type) using the GAPP score and revealed a higher number of CD31 positive vessels in low-grade (well-differentiated) tumors than in moderately differentiated PPGLs. These articles explained the relationship between the vascular pattern and the Zellballen structure, in which fewer vascular vessels were related to the irregular Zellballen structure. S100-positive sustentacular cells are also components of the Zellballen structure and are decreased in metastatic PCCs [44]. In the GAPP grading, low-grade PPGLs had a histology similar to that of the normal adrenal medulla/paraganglia, and the higher grade (moderate to high grades) PPGLs had abnormal structures, an irregular Zellballen structure due to an irregular vascular system, and markedly decreased S100 positive sustentacular cells. Average VHL-related PPGLs were histologically classified into well-differentiated types, indicating that VHL-PPGLs had smaller, regular Zellballen structures surrounded by abundant S100 sustentacular cells and CD31/aSMA-positive small vascular vessels. In contrast, SDHB-related PPGLs had irregular, larger Zellballen structures, with a smaller number of S100 positive sustentacular cells than VHL-related PPGLs. The relationship between genomic mutations and histological characteristics was first described in the present analysis.

Tumor size and metastasis

Our study included 17 PPGLs >5 cm in diameter, of which 10 PCCs and 7 PGLs had 10 VHL, 3 SDHB, 1 MAX, and 3 unknown gene mutations. The largest diameter >5 cm was considered an indicator of metastasis in general PPGLs [45]; however, in pediatric PPGLs, the tumor size itself was not an independent indicator of metastasis at the time of diagnosis. A long-time survey is necessary to establish whether patients with larger tumors have metastasis in the future.

PPGLs with unknown gene mutations

There were three PPGLs with unknown genetic mutations. One patient (case 9) who had no mutations in SDHx and VHL, had been suffering polycythemia and hypertension for the last 3 years. ERAS1, EGLN2, and EGLN1 [46-49] are known gene mutations for PPGL-related polycythemia and may be worth examining for the above patient etiology.

Multiple occurrences and metastasis

Differential diagnosis between multiple occurrences and metastatic tumors should be carefully performed, especially in hereditary PPGLs given that they are usually in multiples and sometimes >5 cm in diameter. The most important factor for patient prognosis is SDHB mutation rather than tumor size, especially in pediatric PPGLs [11]. Tumor metastasis was observed in 2 of 28 patients, which was limited to SDHB-related PGLs, but none of the VHL-related PPGLs in our pediatric series. Patients with metastasis had truncated gene mutations in one patient and missense mutations in the other. It has been reported that truncated mutations have a worse prognosis than missense mutations [50]. A precise diagnosis based on characteristic gross findings, histological findings combined with IHC, and genetic mutations are necessary for patient diagnosis, treatment, and prognosis.

Gene mutation, metastasis, and SSTR2A

Somatostatin and its analogs inhibit the growth of various endocrine and exocrine tumors via SSTR2. The SSTR2A expression in various types of neuroendocrine tumors, including metastatic PCCs, demonstrated strong immunoreactivity of SSTR2A on the cell membranes of tumors, although the types of gene mutations were not identified [23]. Elston et al. [51] reported that 32 tumors in 182 PPGLs were SDH-deficient by using immunohistochemistry, and SDH-deficient tumors were more likely to stain moderately or strongly for SSTR2A when compared to SDH-sufficient tumors (91% vs. 49%, p < 0.0001, respectively). SSTR2 IHC positivity may be a biomarker of potential metastatic behavior of PPGLs in correlation with SDHB-mutation. Indeed, it could be hypothesized that SSTR2 agonists, as shown in neuroendocrine tumor, also decrease HIF-α content in PPGL cells and are therefore especially effective in PPGLs with the disruption of the Krebs cycle, such as SDHx-related PPGLs. Fischer et al. reported the effect of SSTR-based therapy in metastatic PPGLs. SDHB mutations were reported to be associated with SSTR2 IHC positivity and intense staining, suggesting that SSTR2 expression is a significant cell surface therapeutic biomarker of PPGLs [18]. In addition to SDHB-related, MAX-related PPGLs strongly expressed SSTR2 in our cases. These PPGLs could be the target for somatostatin analogue therapy in the case of metastasis or inadequate state for operation, although VHL-related pediatric PPGLs may not be the target for the somatostatin analog, especially in SSTR-IHC negative cases.

Conclusions

To supplement the gene analysis, we investigated IHC for CA9, SDHB, MAX, and clarified that 25 of 28 (90%) pediatric PPGLs had genomic mutations including 14 cases (50%) of VHL, 10 cases (36%) of SDHB, and 1 case (4%) of MAX. The relationship between genomic mutations and histopathological features, GAPP score and SSTR2 expression were comprehensively clarified. Only 2 patients (7%) had metastasis; long time follow-up is necessary to understand whether pediatric patients have a truly low incidence of metastasis, even though 36% of patients had SDHB mutation.

Disclosure

Conflict of Interest

The authors have no conflicts to disclose.

Informed Consent Statement

The work described has been carried out in accordance with The Code of Ethics of the World Medical Association (Declaration of Helsinki) for experiments involving humans. Informed consent was obtained from all the parents of for each child in each hospital prior to genetic testing. Patients older than junior high school signed an informed assent before genetic testing.

Institutional Review Board (IRB) Statement

Muroya K, Sato T received IRB approval from Kanagawa Children’s Medical Center permission for “Clinical and Molecular Genetic Analysis of Pediatric Pheochromocytoma/Paraganglioma for Pathogenesis Elucidation” (# KCMC106-6). Takekoshi K and Yonamine M received IRB approval from the University of Tsukuba Hospital (approval #H28–134; date of approval: September 29, 2019). Kimura N received IRB approval from National Hospital Organization Hakodate Medical Center permission for “Clinicopathological and genomic analysis of pediatric pheochromocytoma and sympathetic paraganglioma.” (#R6-0515001).

Author Contributions

N.K. and K.M. performed study concept and design; K.M., M.Y., T.K., T.S. and N.K. performed development of methodology; K.M., M.Y., K.T., R.H., T.S., K.T., H.M., J.K., H.K., Y.N., and T.K. provided material support, analysis and interpretation of data, and N.K. statistical analysis, and N.K., M.K. and M.Y. writing, review and revision of the paper. All authors read and approved the final paper.

Editorial Board

Takuyuki Katabami is a member of Endocrine Journal’s Editorial Board.