Pheochromocytoma and Paraganglioma Patients With Poor Survival Often Show Brown Adipose Tissue Activation

Abstract

Context

Pheochromocytomas/paragangliomas (PPGLs) are neuroendocrine tumors that can secrete norepinephrine (NE). Brown adipose tissue (BAT) activation is mediated through the action of NE on β-adrenoceptors (β-ARs). In some malignancies, BAT activation is associated with higher cancer activity.

Objective

To study the relationship between BAT activation and PPGL clinical outcomes.

Design

A retrospective case-control study that included 342 patients with PPGLs who underwent ¹⁸F-fluoro-2-deoxy-D-glucose positron emission tomography-computed tomography (¹⁸F-FDG PET/CT) imaging at the National Institutes of Health (NIH). We excluded all patients with parasympathetic tumors and those who underwent ¹⁸F-FDG PET/CT after PPGL resection. Scans of 205 patients were reviewed by 2 blinded nuclear medicine physicians; 16 patients had BAT activation on ¹⁸F-FDG PET/CT [7.80%; age 27.50 (15.00–45.50) years; 10 female/6 male; body mass index [BMI] 24.90 [19.60–25.35] kg/m²). From the remaining 189 patients, we selected 36 matched controls (age 34.4 [25.4–45.5] years; 21 female/15 male; BMI 25.0 [22.0–26.0] kg/m²).

Primary Outcome Measure

Overall survival.

Results

The presence of active BAT on ¹⁸F-FDG PET/CT was associated with decreased overall survival when compared with the control group (HRz 5.80; 95% CI, 1.05–32.05; P = 0.02). This association remained significant after adjusting for the SDHB mutation. Median plasma NE in the BAT group was higher than the control group [4.65 vs 0.55 times above the upper limit of normal; P < 0.01]. There was a significant association between higher plasma NE levels and mortality in PPGLs in both groups.

Conclusions

Our findings suggest that the detection of BAT activity in PPGL patients is associated with higher mortality. We suggest that BAT activation could either be reflecting or contributing to a state of increased host stress that may predict poor outcome in metastatic PPGL.

Pheochromocytomas and paragangliomas (PPGLs) are rare catecholamine-producing neuroendocrine tumors arising from the adrenal medulla and extra-adrenal ganglia, respectively (1). Several studies have shown an increase in brown adipose tissue (BAT) activation in norepinephrine (NE)-secreting PPGLs, ranging from 22% to 30% of patients (2, 3). Up to 70% of patients with PPGLs have germline or somatic mutations in a known susceptibility gene (4). Hereditary PPGLs are divided into 2 main clusters based on genome-wide transcription profiles and clinical behavior. Cluster 1 includes succinate dehydrogenase subunits A, B, C, and D (SDHx), assembly-factor 2 (SDHAF2), and von Hippel-Lindau (VHL). PPGLs that have the noradrenergic biochemical phenotype, such as those found in Cluster 1 disease, have a high frequency of metastatic disease. Cluster 2 primarily includes neurofibromatosis-type 1 (NF1) and rearranged-during-transfection (RET) PPGLs that present with the adrenergic phenotype and rarely with metastatic disease (5). Overall, metastatic PPGLs can occur in about 10% of pheochromocytoma and 20% of paraganglioma patients, with a 5-year survival rate of approximately 50% (5). Although BAT activation is expected to be higher in Cluster 1 patients due to the PPGL-related noradrenergic biochemical phenotype, a recent study found no such association (3).

Unlike white adipose tissue (WAT) whose function is energy storage, BAT is rich in mitochondria and uses glucose and fatty acids for energy expenditure (6). In adults, BAT is located predominantly in the fascial planes in the ventral neck and thorax bilaterally (7). Its unique uncoupling protein 1 (UCP-1) uncouples cellular respiration at the mitochondrial level and dispenses energy in the form of heat instead of adenosine triphosphate (ATP) production. Cold exposure in humans is a well-studied method to activate BAT and make it visible on ¹⁸F-fluoro-2-deoxy-D-glucose positron emission tomography-computed tomography (¹⁸F-FDG PET/CT) (7, 8). This process is mediated by NE-induced β-adrenoceptor (β-AR) stimulation, which not only leads to BAT activation but also its proliferation and differentiation (9–15). BAT activation is associated with multiple metabolic benefits in obesity and metabolic syndrome (16, 17). However, a growing body of evidence has shed light on the association between higher BAT activation in cancer and worse outcomes that are related to increased energy-wasting and cachexia (18–20). Moreover, BAT activity has been recently found to be higher in patients with advanced cancer, suggesting a possible role of BAT in cancer activity (21). However, the effect of BAT activation on the prognosis of patients with PPGL has not been described.

The primary objective of this study was to examine the relationship between BAT activation detected by ¹⁸F-FDG PET/CT and clinical behavior of PPGLs in the background of their specific genotype.

Methods

Patient population

This is a retrospective case-control study that included 342 consecutive patients with genetically characterized and histopathologically confirmed PPGLs who underwent 561 scans by ¹⁸F-FDG PET/CT imaging for disease surveillance at the National Institutes of Health (NIH) between 2013 and 2017 under the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) protocol 00-CH-0093 (ClinicalTrials.gov Identifier: NCT00004847). We extracted data from patient charts: demographic information, body mass index (BMI), clinical signs and symptoms, medications, plasma catecholamines, metanephrines and chromogranin A (CgA), tumor size and location, and clinical outcome, including death and time of the last follow-up. We only excluded patients with parasympathetic paragangliomas and those who underwent ¹⁸F-FDG PET/CT after PPGL resection. Two independent, blinded nuclear medicine physicians reviewed the ¹⁸F-FDG PET/CT scans from 205 patients for evidence of BAT activation. Sixteen patients had BAT activation detected on ¹⁸F-FDG PET/CT (7.80%; age 27.50 [15.00–45.50] years; 10 female/6 male; BMI 24.90 [19.60–25.35] kg/m²). From the remaining 189 patients, using individual-matching, we selected 36 matched controls based on age, sex, and BMI (age 34.40 [25.40–45.50] years; 21 female/15 male; BMI 25.00 [22.00–26.00] kg/m²) (Fig. 1). All patients had their primary tumor resected. We defined high tumor burden as the presence of 10 or more PPGL lesions (22). All patients underwent genetic testing for mutations and deletions in SDHA, SDHB, SDHC, SDHD, SDHAF2, fumarate hydratase (FH), RET, VHL, MYC-associated-factor X (MAX), transmembrane-protein 127 (TMEM127), hypoxia-inducible factor 2A (HIF2A), and prolyl-hydroxylase type 1 and 2 (PHD 1/2). NF1 was diagnosed based on typical clinical presentation. We reviewed death certificates of all patients who died, the cause of death in both groups was attributed to tumor progression, not cardiovascular-related complications. All patients consented in the use of their clinical data, and the Eunice Kennedy Shriver NICHD Institutional Review Board approved the protocol.

Figure 1.

We evaluated 343 consecutive patients with PPGLs who underwent ¹⁸F-FDG PET/CT studies. We excluded patients with normal biochemical profiles as we wanted to study BAT activation and clinical behavior of PPGLs. After exclusion, we had a study cohort of 205 patients with PPGLs who underwent ¹⁸F-FDG PET/CT, among whom 16 patients had evidence of BAT.

Biochemical analysis

We used laboratory values as the primary indicator of remission and recurrence and considered any biochemical elevation above the upper limit of normal (ULN) as evidence of disease. We used 6 laboratory biomarkers as disease surrogates: plasma chromogranin A (CgA, ULN: 93 ng/mL), metanephrine (MN, ULN: 61 pg/mL), normetanephrine (NMN, ULN: 112 pg/mL), epinephrine (E, ULN: 50 pg/mL), NE (ULN: 750 pg/mL), and dopamine (DA, ULN: 29 pg/mL) measured by reversed-phase liquid-chromatographic method, with electrochemical detection as previously described (23). We instructed patients to discontinue medications that could potentially interfere with the laboratory results for 3 weeks, as previously described (23). We included only patients with laboratory values drawn within 1 to 2 weeks of the ¹⁸F-FDG PET/CT scans in the study.

Radiologic analysis and imaging techniques

Detection of BAT was done by ¹⁸F-FDG PET/CT imaging under a previously approved protocol (24). After checking fasting blood sugar level and the injection of ¹⁸F-FDG (7–10 mCi), patients sat for 60 minutes in a quiet room where we set the temperature at 23ºC. Two nuclear medicine physicians blinded to the study outcome interpreted the images. Adipose tissue was identified anatomically and by its density on CT images (−300 to −10 Hounsfield units). Sites of BAT activation assessed were cervical, supraclavicular, axillary, mediastinal, pericardial, and perirenal depots. We reported BAT activation when there was increased uptake of ¹⁸F-FDG with a maximum standardized uptake value > 1.5 in the sites mentioned above (3).

Statistical analysis

Baseline and demographic characteristics were summarized by standard descriptive summaries (eg, means and standard deviations or median and interquartile range for continuous variables and percentages for categorical variables). Paired t-tests were used to compare data that were normally distributed; for nonparametric data, Wilcoxon matched-pairs tests were used. We used the chi-square test, Fisher exact test, or Clopper-Pearson exact binomial test to compare categorical variables. We performed unadjusted Kaplan-Meier survival analysis to test the association between the BAT activation and time-to-death (number of years from the initial PPGL diagnosis till death occurrence). Since the length of disease before the scan can bias the survival analysis interpretation, we compared the time from diagnoses to the time of PET-CT scan in both groups and found no difference between the 2 groups (3.5 [1–7.5] years vs 4.5 [1–10.5] years; P = 0.54]. Time-to-metastasis and time-to-recurrence were defined as the time from diagnosis until the appearance of metastatic lesions in bone and lymph nodes and biochemical recurrence, respectively. The comparison of survival curves was made using the log-rank (Mantel-Cox) test. To account for factors such as NE and SDHB, a Cox proportional hazards model was used. Estimated hazard ratios (HRz) with corresponding 95% confidence intervals (CI) were reported. All analyses were conducted using the statistical software SAS 9.4 (SAS Institute) and GraphPad Prism version 7.01 for Windows (GraphPad Software Inc) with 2-tailed tests based on a significance level of 0.05.

Results

The baseline characteristics of all 52 patients are shown in Table 1. We identified 16 patients with PPGLs who had detectable BAT activity via ¹⁸F-FDG PET/CT (7.80%; age 27.50 [15.00–45.50] years; 10 female/6 male; BMI 24.90 [19.60–25.35] kg/m²) and 36 controls who had no active BAT on ¹⁸F-FDG PET/CT (age 34.40 [25.40–45.50] years; 21 female/15 male; BMI 25.00 [22.00–26.00] kg/m²). Genetically, 63% (10/16) of the BAT group had a gene mutation related to Cluster 1 (Fig. 2A), more than half of which were in SDHx [70% (7/10)] (Fig. 2B). Similarly, the control group had a Cluster 1 predominance [56% (20/36), P = 0.25] with a high SDHx mutations [55% (11/20) P = 0.43] (Fig. 2C, D). We found no significant difference between the BAT and control groups when we compared tumor size, tumor burden, and tumor location (Table 1). Both groups received comparable treatments for metastatic PPGL including chemotherapy with cyclophosphamide, vincristine, and dacarbazine (CVD); tyrosine kinase inhibitors (TKI), temozolomide (TMZ), localized radiation and/or ¹³¹I-meta-iodobenzylguanidine therapy (¹³¹I-MIBG) (Table 2)

Table 1.

Clinical Characteristics in Patients With and Without BAT Detected on¹⁸F-FDG/PET

Patient Characteristics	Controls (36)	BAT Group (16)	P Value
Age, y	34.40 (25.40–45.20)	27.50 (15.00–45.50)	0.20
Sex, female, % (n)	58.33 (21)	62.50 (10)	0.77
BMI, kg/m2	25.00 (22.00–26.00)	24.09 (19.60–25.35)	0.78
Medications
α-AR blockers, % (n)	50.00 (18)	50.00 (8)	0.83
β-AR blockers, % (n)	38.90 (14)	18.80 (3)	0.15
Vital signs
Temperature, °C	36.70 ± 0.50	36.60 ± 0.40	0.25
Mean arterial pressure, mmHg	88.00 ± 9.50	91.90 ± 15.20	0.61
Heart rate, beats/min	79.60 ± 16.80	92.00 ± 13.90	0.01
Biochemical profile
Plasma metanephrine, pg/mL (12–61)	33.50 (17.80–53.00)	31.00 (20.00–47.00)	0.78
Plasma normetanephrine, pg/mL (18–112)	154.00 (66.00–960.00)	959.50 (286.50–4981.00)	0.01
Plasma norepinephrine, pg/mL (112–750)	1415.00 ± 2587.52	6693.13 ± 7753.43	0.01
Plasma Chromogranin A, ng/mL (<93)	1447.14 ± 3897.42	837.69 ± 903.51	0.20
Tumor location and burden
Adrenal % (n)	27.80 (10)	43.80 (7)	0.26
Extra-adrenal PPGL % (n)	41.70 (15.00)	43.80 (7)	0.89
Liver % (n)	19.40 (7.00)	12.50 (2)	0.54
Lung % (n)	25.00 (9.00)	18.80 (3)	0.70
Skeletal % (n)	38.90 (14.00)	50.00 (8)	1.00
Lymph node % (n)	27.80 (10.00)	31.30 (5)	0.88
Patients with high tumor burden# % (n)	33.33 (12)	31.25 (5)	0.72
Size of the largest tumor, cm	3.70 ± 2.60	3.32 ± 2.22	0.69

Data are presented as mean ± SD and median (IQR) for continuous variables and % (n) for categorical variables. Paired t-tests were used to compare data that were normally distributed; for nonparametric data, Wilcoxon matched-pairs tests were used. Comparison of categorical variables was made via chi-square test. Abbreviations: AR, adrenoceptor; BAT, brown adipose tissue; BMI, body mass index; ¹⁸F-FDG PET/CT, 18F-fluoro-2-deoxy-D-glucose positron emission tomography-computed tomography; IQR, interquartile range; PPGL, pheochromocytoma/paraganglioma; SD, standard deviation.

# High tumor burden defined as 10 or more tumors.

Figure 2.

Bar graph of gene clusters (a) and stratification (b) in BAT group and gene clusters (c) and stratification (d) in the BAT group vs control group. Cluster 1 (germ-line variants of succinate dehydrogenase subunits A, B, C, and D and assembly factor 2 [SDHA/B/C/D/AF2], and von Hippel-Lindau [VHL]). Cluster 2 (variants of neurofibromatosis type 1 [NF1], rearranged during transfection [RET] genes), hypoxia-inducible factor 2α (HIF2A).

Table 2.

Different Treatment Modalities Used for Metastatic PPGL in the BAT and Control Groups

Treatment Modalities	Control (36)	BAT (16)	P
Any treatment,* n(%)	27 (75)	10 (62.50)	0.36
CVD, n (%)	12 (33.33)	6 (37.50)	0.76
Local radiation, n (%)	22 (61.11)	9 (56.25)	0.77
Local radiation and CVD, n (%)	9 (25.00)	5 (31.25)	0.74
TKI,# n (%)	13 (36.11)	2 (12.50)	0.10
Local Radiation, CVD and TKI,# n (%)	8 (22.22)	1 (6.25)	0.24
Temozolomide, n (%)	4 (11.11)	2 (12.50)	0.88
131I-MIBG therapy, n (%)	18 (50.00)	3 (18.75)	0.06

Fisher exact test was used to compare groups.

Abbreviations: BAT, brown adipose tissue; CVD, cyclophosphamide, vincristine, and dacarbazine; ¹³¹I-MIBG: Iodine-131-labeled meta-iodobenzylguanidine; TKI, tyrosine kinase inhibitors.

*Any treatment; chemotherapy, local radiation, or 131I-MIBG

^#Axitinib, sunitinib, or imatinib

Time-to-death using an unadjusted Kaplan-Meier survival analysis showed that the presence of active BAT on ¹⁸F-FDG PET/CT was associated with decreased overall survival when compared with the control group (HRz 5.80, 95% CI, 1.05–32.05; P = 0.02), indicating an estimated 5.8-fold increased chance of death in the BAT group compared with controls. This association remained significant after adjusting for the SDHB mutation as a covariant using Cox proportional hazards model (P = 0.04), with an estimated median survival of 11.00 years. (Fig. 3A). All deaths (4/4) in the BAT group were found in patients with SDHx mutations (Fig. 3B).

Figure 3.

Unadjusted Kaplan-Meier survival curves of the BAT vs control group (a) showing a significantly lower overall survival, with a median survival of 11 years. The hazard ratio of death in the BAT group is 5.80; 95% CI, 1.05–32.05; P = 0.02. Unadjusted Kaplan-Meier survival curves of the combination of BAT and SDHx mutation vs SDHx alone (b) showing that the combination of BAT and SDHx is associated with significantly lower survival compared with SDHx alone with P < 0.001. Comparison of survival curves was made using the log-rank (Mantel-Cox) test. The graph does not plot actual raw-data survival proportions, but estimated proportions.

All subjects in the BAT and control groups had biochemically active disease. At the time of the imaging both the BAT and control groups were being treated with a β1-selective-AR antagonist (atenolol 12.5–50 mg/day, 3/16 [18.8%] vs 14/36 [38.9%]) and α-AR antagonists (phenoxybenzamine and/or doxazosin 8/16 [50%] vs18/36 [50%]). In addition, none of the patients in this study were taking labetalol or Demser (metyrosine) at the time of the ¹⁸F-FDG-PET/CT imaging. Patients in both groups had similar levels of thyrotropin (TSH, thyroid stimulating hormone) levels (1.6 miIU/mL [0.92–1.90] vs 2.2 miIU/mL [1.1–3.0], respectively; P = 0.07; 95% CI, −1.26 to 0.07).

Unlike plasma DA and E levels (Fig.4A and 4B), plasma NE levels were significantly higher in the BAT group compared with the control group (4.65 vs 0.55 × median ULN; P < 0.01) (Fig. 4C). This difference in plasma NE levels did not result in any difference in mean arterial blood pressure or body temperature between the 2 groups (92 ± 15 mmHg vs 88 ± 10 mmHg; P = 0.61 vs 36.6 ± 0.3°C vs 36.7 ± 0.5°C; P = 0.25, respectively) (Fig. 4D and 4E). Although there was a higher heart rate (HR) in the BAT group compared with the control group (92 ± 14 bpm vs 80 ± 17 bpm; P = 0.01) (Fig. 4F), using a general linear model, there was no group difference in HR (P = 0.50) after taking NE into the effect. There was no significant association between HR and time-to-death between the BAT and control groups. A proportional hazard model analysis with plasma NE as a covariate identified a significant association between higher plasma NE levels and mortality in PPGLs in both groups (HRz = 1.09, 95% CI, 1.02–1.153; P < 0.01). The effect of BAT and NE combined was significant in decreasing survival in the BAT group compared with the control group (HRz = 6.26, 95% CI, 1.12–35.00; P = 0.03), and after adjusting this combined analysis to plasma NMN, we found that the presence of BAT was still significantly associated with increased mortality (HRz 5.6; CI, 1.002–31.54; P = 0.04). However after adjusting to plasma NE levels, the effect of BAT and NE combined on mortality in PPGLs was no longer significant (HRz = 1.06; 95% CI, 0.90–1.18; P = 0.14). All deaths in the BAT group belonged to the upper 75^th percentile of the NE range. Due to sample size, we were not able to perform a subgroup analysis to identify the effect of BAT alone on mortality. That is, a larger number of patients will need to be studied to determine more definitively whether there is an association between BAT activation and to time-to-death.

Figure 4.

The difference in median (IQR) of plasma dopamine (a), epinephrine (b) and norepinephrine (NE) (c) showing significantly higher NE in the BAT group compared with control (c). Mean ± SD of mean arterial pressure in mmHg (d), body temperature (°C) (e), and heart rate (beats/min) (f) in the BAT group compared with control group, respectively. Values are presented as multiples of the ULN, respectively. Paired t-tests were used to analyze data that were normally distributed; for nonparametric data, Wilcoxon matched-pairs tests were used. Abbreviation: ULN, upper limit of normal.

We compared mortality in the BAT and control groups and found that the patients in the BAT group died younger (23.0 [17.2–33.5] vs 48.0 [46.5–49.8] years) and had markedly elevated plasma NE (22.10 ± 8.00 vs 0.48 ± 0.27 × ULN) (Table 3). In an unadjusted Kaplan-Meier survival analysis, we found no difference in time-to-metastasis (P = 0.60) and time-to-recurrence between the BAT and control groups (P = 0.25).

Table 3.

Clinical Characteristics of Mortality in the BAT Group and Control Group

Patient Characteristics	Control Group (2)	BAT Group (4)
Age, y	48 (46.55–49.80)	23 (17.20–33.50)
Sex, females, n (%)	1 (50.00)	1 (25.00)
BMI, kg/m2	26.0 ± 1.41	25.9 ± 1.60
Vital signs
Temperature, ◦C	37.00 ± 1.84	36.25 ± 0.33
Systolic blood pressure, mmHg	122.00 ± 12.73	145.75 ± 22.02
Heart rate beats/min	100.00 ± 24.04	100.50 ± 12.71
Biochemical profile
Plasma metanephrine, pg/mL (12–61)	31.00 ± 11.31	37.50 ± 17.41
Plasma normetanephrine, pg/mL (18–112)	103.00 ± 74.95	9747.75 ± 5790.09
Chromogranin A, ng/mL (<93)	10 552.00 ± 14 521.14	1385.25 ± 156.17
Plasma norepinephrine, pg/mL (112–750)	371.00 ± 190.91	17 301.00 ± 6305.30
Plasma norepinephrine x ULN*	0.48 ± 0.27	22.12 ± 8.02
Tumor location and burden
Adrenal n (%)	1 (50.00)	1 (25.00)
Extra-adrenal PPGL n (%)	2 (100.00)	2 (50.00)
Patients with high tumor burden# n (%)	2 (100.00)	3 (75.00)
Size of the largest tumor, cm	7.25 ± 2.47	2.85 ± 1.48

Data are presented as mean ± SD and median (IQR) for continuous variables and n (%) for categorical variables. P values not shown due to the small sample size.

*Plasma norepinephrine is shown as mean multiples of plasma NE upper limit of normal

#High tumor burden defined as ten or more tumors

Discussion

The objective of our study was to determine if the BAT activation detected by ¹⁸F-FDG PET/CT in PPGL patients would have any prognostic clinical implications. We found not only that the detection of BAT activity by ¹⁸F-FDG PET/CT in PPGL patients was associated with decreased overall survival, but also that elevated NE was an independent risk factor for mortality likely unrelated to cardiovascular complications, especially in patients with an SDHx variant. Since NE is a known BAT activator (6–10), BAT activation detected on ¹⁸F-FDG PET/CT could be a biomarker of prolonged increased sympathetic tone and/or other humoral factors associated with adverse outcome in metastatic PPGL.

Role of sympathetic tone in metastatic PPGL and cancer prognosis

There is a growing interest in identifying prognostic markers that can predict metastatic disease and survival in PPGL patients. A recently published meta-analysis evaluated the power of genotype-phenotype correlations in predicting outcome in 21 studies comprising 703 PPGL patients (25). There was a correlation between the SDHB mutation as well as elevated NE and DA levels with metastasis occurrence (25). Another recent retrospective multicenter study examined prognostic factors that determined survival in patients with metastatic PPGL and demonstrated that better outcomes were associated with head and neck tumors, age < 40 years, low proliferative index, and metanephrines of < 5 times the ULN (26). Based on these studies and our current work, we may conclude that the NE level is a predictor of a negative outcome in PPGL patients.

Not only PPGLs, but also other tumors seem to be sensitive to the changes in sympathetic tone and subsequent stimulation of β-ARs. This observation was first identified in epidemiological studies showing that chronic use of β-AR antagonists was associated with lower recurrence and mortality of breast cancer, melanoma, and prostate cancer (27). A subsequent series of preclinical and clinical studies showed that increased sympathetic tone (eg, mental or physical stress and use of drugs that inhibit catecholamine clearance) predispose to various cancer types through the activation of oncogenic signaling downstream of the β-AR (27, 28). Although increased sympathetic tone might not always be measured in cancer patients, it can be identified by the presence of active BAT on ¹⁸F-FDG PET/CT scans that are routinely done as part of disease surveillance (7, 8).

BAT activation and cancer outcomes

BAT is a metabolically active tissue that has recently been recognized in adult humans to have a role in energy expenditure (29). Cold-induced, NE-mediated β-AR stimulation is a well-established method to activate BAT and make it detectable on ¹⁸F-FDG PET/CT (16, 30). Since increased sympathetic outflow predicts adverse outcome in malignant tumors, including PPGL, and BAT activation correlates with NE levels (3), BAT activity detected ¹⁸F-FDG PET/CT could be a biomarker for outcome in PPGL patients as for other cancers. There is a growing body of evidence that has shown an association between BAT activation and cancer prognosis (21). A study in mice showed accelerated tumor growth following tumor implantation into BAT (31).

In humans, BAT activity was associated with breast cancer progression (32) and higher overall cancer activity (21). Some attribute poor outcomes observed in BAT activation to cachexia and wasting mediated by BAT increased energy expenditure (18–20). However, this mechanism was not supported by others. In a large study that evaluated the association of BAT activity, ¹⁸F-FDG PET/CT results, and cancer tumor activity, patients with active cancer had significantly higher BAT volume compared with patients without active disease (21). In some studies, BAT activity could reflect a state of an increased sympathetic tone mediated by the cancer-induced stress response, triggering an oncogenic signal (27, 28). Alternatively, BAT could contribute to decreased host tolerance by inducing a negative energy balance (18–20), although it remains to be determined how the total mass of active BAT affects the prognosis of cancer patients, including their weight changes. In our study, patients with detectable BAT had the highest mortality among those with metastatic disease and those who had the highest levels of NE, up to 22 times the ULN. Even though both groups had a similar BMI, the “normal” BMI in the BAT group may represent wasting of muscle tissue and not truly a reflection of a leaner, healthier metabolic profile. Therefore, BAT activation in metastatic PPGL, like other cancers, may reflect a pathological increase in sympathetic tone that reflects aggressive disease, host stress, and poor outcome.

BAT activation beyond sympathetic stimulation

Although we have shown that NE was significantly higher in the BAT group, 25% of patients in BAT activation group not control group had a NE within the reference range at thermoneutrality. This raises the question of whether there are other “browning” factors in PPGL. BAT activation and/or WAT browning can be induced by multiple factors including thyroxine, bile acids, interleukin 6 (IL-6), parathyroid hormone-related protein (PTHrP), fibroblast growth factor 21, and natriuretic peptides (20). IL-6 and PTHrP have been identified as inducers of browning in mice, even at thermoneutrality in cancer cachexia and humans (18, 20, 33). PPGLs are known to secrete catecholamines, but they can also secrete various growth factors that are responsible for some of the autocrine or paracrine functions (34).

Moreover, SDHx germline mutations predispose PPGLs to accumulate succinate (35). Accumulation of succinate is believed to make Cluster 1 tumors more avid on ¹⁸F-FDG PET/CT (1). Succinate can theoretically leak into circulation (35) and be sequestrated in BAT (36). If succinate also accumulates in BAT, then a higher ¹⁸F-FDG PET/CT avidity of BAT might be expected in SDHx-related PPGL. Since hereditary PPGLs have an autosomal dominant inheritance, succinate may get oxidized by the residual SDH enzyme activity and induce UCP-1 thermogenesis (36). Therefore, more studies are needed to uncover any association between metastatic PPGLs and browning factors other than NE.

We observed a lower incidence of BAT activation (8%) in the PPGL patients in our study when compared with other PPGL cohorts (20%–30%) (2, 3). This lower rate could be related to the ambient temperatures in the rooms where patients sit for 60 minutes while the FDG distributes to the tissues. Unlike other studies that reported ¹⁸F-FDG PET/CT conduction at room temperatures as low as 20°C (18), our ¹⁸F-FDG PET/CT suite is kept deliberately warmer, at 23°C, which substantially reduces the confounding effect of cold-induced BAT activation. The use of nonselective β-AR blockers, such as propranolol, has been described to cause a reduction in BAT activation in PPGLs (3, 8, 37). This aspect was not applicable in our patient cohort since the use of both α and β1-selective-AR blockers were similar in both groups.

Our study has several limitations. Like any retrospective observational study design, several factors could not be controlled. Outside temperature is an essential factor in BAT activation (38). The effect of outside temperature was not considered in this study; however, since we included only PPGL patients who had biochemically active disease and had their scans done after resting at thermoneutrality for 60 minutes, the seasonal weather variations were likely minimized. Our BAT group sample size limited further subgroup mortality analysis between patients with high and low NE levels, and therefore, the effect of BAT activation and NE excess could not be separated. Moreover, we did not perform detailed quantification of BAT, due to the potential bias by false-positive signals from tumors on ¹⁸F-FDG PET/CT in metastatic PPGL that were abundant in our cohort.

In conclusion, our findings suggest that the detection of BAT activity on ¹⁸F-FDG PET/CT in PPGL patients, especially with SDHx mutation is associated with higher mortality. The exact mechanism of this association is not clear. However, we suggest that BAT activation in metastatic PPGL could either be reflecting or contributing to a state of increased host stress that may predict poor outcome. Future studies should focus on the mechanisms by which NE-mediated BAT activity affects survival among PPGL patients.

Glossary

Abbreviations

¹⁸F-FDG PET/CT

¹⁸F-fluoro-2-deoxy-D-glucose positron emission tomography-computed tomography

ATP

adenosine triphosphate

β-AR

β-adrenoceptor

BAT

brown adipose tissue

BMI

body mass index

CgA

chromogranin A

CVD

cyclophosphamide, vincristine, and dacarbazine

DA

dopamine

E

epinephrine

HR

heart rate

HRz

hazard ratio

MN

metanephrine

NE

norepinephrine

NIH

National Institutes of Health

NMN

normetanephrine

PPGL

pheochromocytomas/paraganglioma

SDH

succinate dehydrogenase

UCP-1

uncoupling protein 1

ULN

upper limit of normal

WAT

white adipose tissue

Additional Information

Disclosure Summary: The authors claim no conflicts of interest.

Data Availability: The datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.