Purpose: Germline mutations in genes encoding mitochondrial succinate dehydrogenase (SDH) are found in patients with paragangliomas, pheochromocytomas, gastrointestinal stromal tumors, and renal cancers. SDH inactivation leads to a massive accumulation of succinate, acting as an oncometabolite and which levels, assessed on surgically resected tissue are a highly specific biomarker of SDHx-mutated tumors. The aim of this study was to address the feasibility of detecting succinate in vivo by magnetic resonance spectroscopy.
Experimental Design: A pulsed proton magnetic resonance spectroscopy (1H-MRS) sequence was developed, optimized, and applied to image nude mice grafted with Sdhb−/− or wild-type chromaffin cells. The method was then applied to patients with paraganglioma carrying (n = 5) or not (n = 4) an SDHx gene mutation. Following surgery, succinate was measured using gas chromatography/mass spectrometry, and SDH protein expression was assessed by immunohistochemistry in resected tumors.
Results: A succinate peak was observed at 2.44 ppm by 1H-MRS in all Sdhb−/−-derived tumors in mice and in all paragangliomas of patients carrying an SDHx gene mutation, but neither in wild-type mouse tumors nor in patients exempt of SDHx mutation. In one patient, 1H-MRS results led to the identification of an unsuspected SDHA gene mutation. In another case, it helped define the pathogenicity of a variant of unknown significance in the SDHB gene.
Conclusions: Detection of succinate by 1H-MRS is a highly specific and sensitive hallmark of SDHx mutations. This noninvasive approach is a simple and robust method allowing in vivo detection of the major biomarker of SDHx-mutated tumors. Clin Cancer Res; 22(5); 1120–9. ©2015 AACR.
A large proportion of patients with paraganglioma/pheochromocytoma (PPGL) carry a germline mutation in an SDHx gene. Identification of SDHx mutations is important for the diagnostic work-up, the management, and the follow-up of patients with PPGL and their families, who are at risk of developing multiple paraganglioma. Moreover, SDHB gene mutations predispose to malignant, particularly aggressive forms of the disease. Therefore, a genetic counseling is now recommended for all patients suffering from PPGL. We here show that in vivo detection of succinate by proton magnetic resonance spectroscopy is a highly specific and sensitive hallmark of SDHx mutated tumors. This noninvasive approach will allow identifying and classifying SDHx mutations or variants of unknown significance. It may help for the characterization of inoperable tumors and suspicious lesions and serve as a surrogate biomarker in the assessment of tumor response to specific treatments.
Pheochromocytoma and paraganglioma are rare neuroendocrine tumors that arise in chromaffin cells of the adrenal medulla and in sympathetic and parasympathetic ganglia, respectively. The prevalence of pheochromocytoma and paraganglioma (PPGL) in patients with hypertension consulting at general outpatient clinics is estimated at 0.2% to 0.6%, but this number may be underestimated (1). Nearly 40% of patients with PPGL carry a germline mutation in one of the 13 PPGL predisposing genes identified so far (2), and mutations of SDHx genes (SDHA, SDHB, SDHC, SDHD, SDHAF2) are causative of approximately half of the genetically determined cases. SDHx mutations predispose to the hereditary PPGL syndrome but may also be found in patients with gastrointestinal stromal tumors (GIST; ref. 3) or renal clear cell carcinomas (4).
SDHA, B, C, and D genes encode the four subunits of succinate dehydrogenase (SDH), a mitochondrial enzyme of the tricarboxylic acid (TCA) cycle that oxidizes succinate into fumarate. They were the first genes encoding a mitochondrial enzyme demonstrated to act as tumor suppressor genes (5), an important finding supporting the hypothesis of a direct link between mitochondrial dysfunction and cancer proposed by Warburg in 1924 (6). Since then, mutations in genes encoding for the TCA enzymes fumarate hydratase (FH; ref. 7), isocitrate dehydrogenase (IDH1 and 2; ref. 8), and more recently malate dehydrogenase (MDH2), were reported to predispose to PPGL, renal cancers, leimyomas, or to be causative of sporadic gliomas (for review, see ref. 9).
Identification of SDHx mutations is important for the diagnostic work-up, the management, and the follow-up of index cases and their families. SDHx mutation carriers are at risk of developing multiple paraganglioma that can arise all along the embryonic migration way of neural crest cells, from the base of the skull to the pelvis (10). Moreover, the identification of SDHB gene mutations is of specific clinical importance as they predispose to malignant, particularly aggressive forms of the disease (11, 12), and a genetic counseling is now recommended for all patients suffering from PPGL (1). In familial PPGL patients carrying a germline heterozygous mutation on an SDHx gene, the somatic loss of the remaining allele induces a complete SDH loss-of-function, which results in the accumulation of succinate. Succinate acts as an oncometabolite and is suspected to mediate most, if not all, of the tumorigenic effects related to SDHx mutations (9, 13, 14). Development of specific tumor biomarkers allowing the rapid identification of these patients would be highly beneficial and particularly helpful for the characterization of inoperable tumors and suspicious lesions. Biomarkers could serve as surrogate markers in the assessment of tumor response to specific treatments. Until now, no in vivo method to assess the functional consequences of SDHx mutations was available, and all existing tests were performed on surgical resected specimens (15–22).
Succinate concentrations in the millimolar range—4 to 500 micromoles per gram depending on studies and procedures—have been reported in SDHx-mutated PPGL tumors, an increase of up to 100-fold compared with non-SDHx–mutated PPGL tumors (19, 21, 23). We hypothesized that these succinate levels could be detected noninvasively by in vivo proton magnetic resonance spectroscopy (1H-MRS) in SDHx-mutated tumors, without the need for tissue sampling, similarly to 2-hydroxyglutarate in patients with IDH1/2-mutated gliomas (24–26). Here, we report a new method for SUCCinate Estimation by Spectroscopy (SUCCES) in patients with PPGL related or not to an SDHx mutation.
Materials and Methods
Optimization of SUCCES for succinate detection with 1H-MRS
We first tested the SUCCES sequence at 4.7 T on a 3-cm-diameter spherical phantom tube containing 100 mmol/L sodium succinate dibasic hexahydrate, 100 mmol/L l-lactate, and animal fat. Spectra and signal intensities acquired with different echo times (TE) ranging from 12.8 to 792 ms are shown in Supplementary Fig. S1A and S1B. The gradual decrease of the lipid signal with increasing TE eventually unmasked the succinate peak at TE > 70 ms and the lactate peak at TE > 100 ms. As previously described (27), increasing TE reduced the succinate signal exponentially and the lactate peak sinusoidally due to scalar coupling effect (Supplementary Fig. S1B). To study both the succinate and lactate signals, we chose a TE of 272 ms that reduced fat contamination and yielded positive lactate and succinate peaks. We then performed 1H-MRS spectra with decreasing (10–1 mmol/L) concentrations of succinate and lactate [repetition time (TR), 3000 ms; echo time (TE), 272 ms; average, 128; volume of interest (VOI) size, 5 × 5 × 5 mm3]. Water suppression was performed using VAPOR pulses (sinc RF pulses, 646 ms total duration, 700 Hz bandwidth), followed by crusher gradients (3 ms duration, 58 mT/m strength). Three-sinc–shaped RF pulses with 4 kHz bandwidth to obtain a spectral width of 20 ppm achieved VOI selection. Spectral resolution was 0.98 Hz/point. After zero filling and phase correction, data filtration was performed with a Gaussian function at the top of the Free Induction Decay, with a length band of 2 Hz and visualization was obtained with a Fourier transform. The area under the succinate peak measured from the spectra was linearly correlated with the succinate concentration (Supplementary Fig. S1C and S1D).
A similar procedure was performed in the 3T clinical scanner using 3-cm-diameter phantom tubes and a larger VOI size (10 × 10 × 10 mm3), except that the TR was 2,500 ms and the TE was lowered to 144 ms to compensate the lower detection threshold and the decrease of the signal-to-noise ratio (Supplementary Fig. S1E). On the basis of these procedures, the threshold for succinate detection was found to be approximately 1 mmol/L in both magnets.
Generation of the allografted mouse model
No animal model of SDH-related PPGL being available, we generated an allografted mouse model by subcutaneous injection of 2.5 × 106 immortalized mouse chromaffin cells (imCC; ref. 13) carrying a homozygous knockout of the Sdhb gene (Sdhb−/−, clone 8) or their wild-type (WT) counterparts (Sdhblox/lox) into the flanks of 10-week-old female NMRI-nu mice. Animal experiments were registered by the French Ethical committee under Number 14-041 and followed the ARRIVE guidelines of the National Centre for the Replacement, Refinement, and Reduction of Animals in Research (London, UK).
Tumors were allowed to grow until their size reached 0.6 cm3. They were then resected and 8 mm3 fragments were grafted in the dorsal fat pad of naïve nude mice. The tumors grew in 100% of mice and tumors were macroscopically visible after 2 weeks for WT and 1 month for Sdhb−/− tumors, in line with the reduced growth rate of Sdhb−/− cells observed in vitro (13). Immediately after magnetic resonance spectroscopy, tumors were retrieved and snap-frozen in liquid nitrogen or fixed in 4% paraformaldehyde.
Succinate 1H-MRS in a 4.7-T animal-dedicated MRI device
Imaging was performed 37 ± 11 days after the graft for the Sdhb−/− group versus 22 ± 7 days for the WT group. Mice were placed in prone position under isoflurane anesthesia (4% for induction and 1.5% for maintenance in 1 L/min air) with respiration monitoring.
1H-MRS was performed in a dedicated small-animal 4.7-T MR system (Biospec 47/40 USR Bruker), using a 1H quadrature transmit/receive body coil with a 3.5-cm inner diameter. An anatomical two-dimensional (2D) steady-state free precession sequence (True FISP) was first acquired in two orthogonal planes. 1H-MRS was then carried out using an optimized asymmetric Point REsolved SpectroScopy (PRESS) monovoxel acquisition (Supplementary Fig. S1). Echo signals [TR, 3,000 ms; TE, 144 or 272 ms; average, 512, with a VOI size of 5 × 5 × 5 mm3) were acquired during 25 minutes.
The MRS spectrum of succinic acid [HOOC-(CH2)2-COOH] presents a characteristic peak at 2.44 ppm, corresponding to the precession frequency of the CH2 protons. The succinate peak was quantified by measuring the area under the peak using Topspin 2.0 software (Brüker Corporation).
Patients were recruited from the French COMETE (‘Cortico et Médullosurrénale: les Tumeurs Endocrines’) cohort of the Hypertension unit of the European Georges Pompidou Hospital (HEGP), Paris, France. Ethical approval for the study was obtained from the institutional review board [Comité de Protection des Personnes (CPP) Ile de France II], and written informed consent to participate in the study was obtained from all patients. The procedures used for PPGL diagnosis and genetic testing were in accordance with international clinical practice guidelines (1). Mutation analysis of PPGL susceptibility genes was performed as previously described (28). When patients underwent surgery for paraganglioma, fresh tumor samples were frozen immediately after surgical removal and stored in liquid nitrogen until processing following the COMETE collection procedures. Confirmation of diagnosis was performed by histology on paraffin-embedded, formalin-fixed samples.
Succinate detection by 1H-MRS in patients at 3 T
Combined MR images and MR spectroscopic scans of patients were acquired in a 3 T MRI clinical scanner (Discovery MR750w GEMSOW, GE Medical Systems). 1H-MRS spectra were acquired by PRESS on the basis of the PROBE monovoxel sequence (29) and optimized for succinate and lactate detection, with TR, 2,500 ms; TE, 144 ms; Nex: 512 (22-minute acquisition) or 1,024 (44-minute acquisition). The VOI (1.3–19 cm3) was centered on the anatomical image to prevent lipid contamination from the tissue surrounding the tumor as previously described (30).
MR images were acquired using a whole body (GEM Chest/Body/Pelvis; Body 24 AA3) or a head and neck (GEM Head/Neck/Chest; Head 24) phased-array multi-coil.
Detection of tumors and VOI positioning was performed on thin-section high-resolution T2-weighted fast spin-echo imaging in at least two orthogonal planes with the following parameters: TR, 2,500 ms; TE, 85 ms; echo train length, 19; slice thickness, 3 mm; spacing, 0.3; field of view, 14 × 14 cm2 for neck or 42 × 42 cm2 for whole-body coil; matrix, 320 × 320.
If necessary, an anatomical 2D steady-state free precession sequence (FIESTA CINE) was acquired with TR, 3.7 ms; TE, 1.4 ms; TI, 210 ms; slice thickness, 5 mm; spacing, 1 mm; field of view, 14 × 14 cm2 for neck or 42 × 42 cm2 for whole-body coil and/or a 3D angio-MR at arterial phase, after contrast agent administration of gadoterate meglumine 0.2 mL/kg with TR, 11.4 ms; TE, 2.2 ms; slice thickness, 0.8 mm; spacing, 0.4 mm; field of view, 30 × 27 cm2.
A prescan algorithm was first acquired to adapt the transmitter and receiver gains and center frequency, the homogeneity of the magnetic field was optimized with the three-plane auto-shim procedure, and water suppression and automatic shimming of the single voxel were performed.
Measurement of succinate by gas chromatography/mass spectrometry
Tumor samples from 15 mice (10 samples with Sdhb knockout and 5 WT controls) and from paragangliomas of four patients (patients #1, #5, #6, and #9) were processed by organic extraction with ethylacetate, derivatization with N,O-bis(trimethylsilyl) trifluoroacetamide with 1% trimethylchlorosilane, and analyzed by gas chromatography/tandem mass spectrometry (GC-MS) on a GC-MS triple quadrupole (Scion TQ, Brüker Daltonics). Analytes were identified according to retention time and mass spectrum in selected reaction monitoring mode on the basis of standard spectral reference libraries.
SDHA, SDHB, and SDHD protein expression were assessed on formalin-fixed, paraffin-embedded (FFPE) tumor samples by immunohistochemistry as previously described (16, 18, 22) using the following antibodies: anti-SDHA (ab14715, Abcam; 1:1,000), anti-SDHB (HPA002868, Sigma-Aldrich Corp; 1:500), and anti-SDHD (HPA045727, Sigma-Aldrich Corp; 1:50).
SDH activity was investigated on frozen tumor samples using a spectrophotometrical assay, as previously described (31).
In vivo detection of succinate in murine allografted tumors
The 1H-MRS sequence was optimized in vitro (Supplementary Fig. S1). To investigate whether in vivo detection of succinate could be assessed for the noninvasive identification of SDH-related tumors, a proof-of-concept pilot study was performed in a mouse model prior to patients' exploration. No animal model of SDH-related PPGL being available, we generated an allografted mouse model using imCC carrying a homozygous knockout of the Sdhb gene or their WT counterparts (13). Sdhb knockout in tumors was confirmed by genotyping (Fig. 1A) and measurement of SDH enzymatic activity (Fig. 1B). GC-MS showed a massive accumulation of succinate in SDH-deficient tumors: 28.3 ± 9.5 nmol per mg protein in the Sdhb−/− group, versus 0.6 ± 0.7 nmol/mg protein in the control groups (Fig. 1C), confirming inhibition of SDH activity in Sdhb−/− tumors. 1H-MRS was first tested in mice using a TE = 272 ms and a fixed VOI size (125 mm3) placed over the tumor mass of 13 Sdhb−/− and 16 WT allografted mice (Fig. 1D). The peak corresponding to lactate, indicative of anaerobic glycolysis, was always present regardless of the tumor type. In contrast, the succinate peak was only detected in Sdhb−/− tumors, with a sensitivity and specificity of 100% (n = 13), in agreement with succinate accumulation caused by SDH inhibition (Fig. 1D; Supplementary Fig. S2A). Measurements of succinate concentrations in Sdhb−/− tumor samples (n = 4) by GC-MS confirmed the MRS results. The succinate levels measured in vitro correlated with the area under the succinate peaks using a TE of 272 ms (r2 = 0.88; Fig. 1E). The results obtained in the 4.7 T magnet with TE = 272 ms were repeated with TE = 144 ms in 5 Sdhb−/− and 3 WT tumors, (Fig. 1F; Supplementary Fig. S2B). At TE = 144 ms, lactate was hardly detectable, whereas succinate was specifically observed in Sdhb−/− tumors. Again, GC-MS quantification of succinate performed in Sdhb−/− resected samples (n = 5) correlated with in vivo measures (r2 = 0.70; Fig. 1G).
In vivo detection of succinate in patients
Nine patients presenting with pheochromocytoma, cervical, and/or abdominal paragangliomas were recruited at the Hypertension unit of the European Georges Pompidou Hospital (Table 1). All patients benefited from genetic counseling in accordance with the Endocrine Society Clinical Practice Guidelines (1). Before undergoing SUCCES with 1H-MRS, a germline SDHx gene mutation was identified in four patients (one SDHB, one SDHC, and two SDHD), whereas no mutation was identified for the other five patients.
|Patient .||Gender .||Age, y .||Gene .||Mutation type .||Multiple locations .||Type of tumor analyzed .||Largest tumor diameter, mm .||VOI size, cm3 .||Succinate level on GC-MS, nmol/mg protein .|
|1||M||33||SDHB||c.740T>Gp.Met247Arg||Yes||VPGL||80||4.6 and 1.5||ND|
|Patient .||Gender .||Age, y .||Gene .||Mutation type .||Multiple locations .||Type of tumor analyzed .||Largest tumor diameter, mm .||VOI size, cm3 .||Succinate level on GC-MS, nmol/mg protein .|
|1||M||33||SDHB||c.740T>Gp.Met247Arg||Yes||VPGL||80||4.6 and 1.5||ND|
Abbreviations: APGL, abdominal paraganglioma; CBPGL, carotid body paraganglioma; NA, not applicable; ND, not determined; PCC, pheochromocytoma; VPGL, vagal paraganglioma.
Genetic testing identified a variant of unknown significance (VUS) in the SDHB gene of Patient #1 (c.740T>G = p.Met247Arg), a 33-year-old male with two paragangliomas and a predominant noradrenergic secretion profile. In the cervical paraganglioma of this patient, the 1H-MRS signal of succinate was unequivocally discernable at 2.44 ppm using 1,024 averages and a 4.6 cm3 VOI size (Fig. 2A). Sensitivity and specificity were explored by reducing the scan repeats from 1,024 to 512 averages and the VOI size from 4.6 to 1.5 cm3: the succinate peak was still clearly detected in low sensitivity conditions in the cervical paraganglioma (Fig. 2A), as well as in the abdominal tumor mass (Fig. 2B), but not in the liver, showing the persistence of SDH activity in this healthy organ, expected to be heterozygous for the mutation (Fig. 2C). The pathogenicity of this newly described variant suggested by 1H-MRS was supported by the LOH at the SDHB locus in DNA extracted from the resected abdominal paraganglioma (Fig. 2D) and confirmed by three functional tests: SDHB-negative (unspecific weak diffuse signal) and SDHD-positive immunohistochemistries (Fig. 2E), loss of SDH enzymatic activity (Fig. 2F), and succinate accumulation measured by GC-MS (Fig. 2G).
The succinate peak was observed in the tumors of the three other SDHx patients using scan repeats of 1,024 and 512 averages (Supplementary Fig. S3). Interestingly, a choline peak at 3.2 ppm was associated with the succinate peak in each of the SDHx-mutated tumors. In contrast, neither succinate nor choline peaks were observed in tumors from patients without SDHx mutations (Fig. 3; Supplementary Fig. S4). In Patient #5, immunohistochemistry of tumor samples showed SDHB- and SDHA-positive and SDHD-negative staining, whereas GC-MS analysis confirmed the absence of succinate accumulation (Fig. 3B and C).
Unexpected SDHA mutation identified by SUCCES
Surprisingly, a small but significant peak above baseline was detected in an abdominal paraganglioma from a patient with an apparently sporadic form of the disease (Fig. 4A.) This patient was a 48-year-old man suffering from a single abdominal paraganglioma with no family history of PPGL. Following comprehensive genetic counseling according to the international guidelines, the search for mutations of SDHB, SDHC, and SDHD genes returned negative. Nevertheless, the presence of a succinate peak in the tumor of this patient prompted us to sequence the SDHA gene of this patient, which identified a c.91C>T = p.Arg31Ter mutation (Fig. 4B), previously reported in Dutch patients with PPGL (16) or GIST (32). After the patient had undergone surgery, additional analyses of his tumor showed negative SDHA and SDHB and positive SDHD immunohistochemistry (Fig. 4C). GC-MS confirmed the accumulation of succinate (Table 1) and validated the rare and unexpected SDHA-mutated status of this patient that had been initially stratified as a sporadic case.
Here, we report the noninvasive detection of succinate by in vivo MRS in tumors of patients with PPGL carrying SDHx genes mutations but not in those of patients without SDHx mutations. Interestingly, in an Sdhb−/− mouse tumor model, this succinate peak is correlated with the concentrations of succinate measured in the resected tumors by GC-MS.
Demonstration of SDH inactivation is currently based on in vitro analyses of tissue samples: immunohistochemical analyses of SDHB, SDHA, and SDHD expression in FFPE tissues (16, 18, 22), direct succinate measurements on frozen tumor samples by nuclear magnetic resonance (NMR) spectroscopy (15, 19, 20, 23), GC-MS, or liquid chromatography mass spectroscopy (LC-MS; refs. 13, 17, 21). Recently, Varoquaux and colleagues reported in vivo detection of succinate using 1H-MRS in 6 patients with head and neck paraganglioma (3 SDHD, 1 SDHB, and 2 sporadic cases). Although the spectra quality was considered as low in the 2 sporadic cases and uninterpretable in one SDHD-mutated tumor, a succinate peak was also only detected in three SDHx-mutated tumors (33). In the present study, we show that 1H-MRS also detects succinate in abdominal paraganglioma and in genes encoding all four SDH subunits. Moreover, we performed longer acquisition time (512 and 1,024 averages, vs. 120 in the Varoquaux and colleagues study), which allowed an immediate interpretation of spectra, without the need of postprocessing the data.
The benefits of assessing this tumor hallmark in patients with SDHx-related tumors are important in several aspects. SUCCES would allow stratifying these patients or classifying VUS as deleterious mutations with no need of tissue sampling. Patient #9 carried a single abdominal paraganglioma diagnosed at age 48, without a family history for this disease. According to international guidelines, SDHB, SDHC, and SDHD genetic testing were performed in this patient, but not SDHA, which would have been prescribed only after surgery, in case of SDHA negative immunohistochemistry (1). In such a case, exploring the patient with 1H-MRS and detecting the succinate peak orientated us without hesitation toward SDHA sequencing, leading to early identification of the mutation. SDHA immunohistochemistry is not included in international guidelines and is not a standard procedure, thus it is likely that this mutation would have been missed in most instances. In Patient #1, an SDHB gene VUS was identified, whereas SDHB immunohistochemistry showed a potentially misleading, weak diffuse signal, previously reported in some PPGL with SDHx genes mutations (18). Hence, in this other case, the 1H-MRS succinate peak was particularly informative to validate the functionality of the SDHB mutation.
SDHx mutation carriers are at risk of developing multiple paragangliomas, and SDHB-mutated carriers are predisposed to metastatic forms of the disease. Knowledge of the SDHx-mutated status is critical for the follow-up and clinical management of patients and of their relatives. On the basis of an early knowledge of the SDHx mutational status, surgeons may decide to adapt their procedures, especially for SDHB cases. For nonoperable tumors, therapeutic choices may also take advantage from this information. For example, studies have suggested that SDHB mutation carriers may be better responders to high doses of 131I-MIBG (34), sunitinib (35), or temozolomide treatments (36). Although these results will need to be evaluated in larger, prospective, and comparative studies, they nevertheless pave the way toward personalized medicine for inherited PPGL.
Overall, the clinical value of SUCCES lies in its capacity to assess for the presence of succinate repeatedly over the time course of the disease, for clinical surveillance, postoperative follow-up, and evaluation of treatment efficacy (17). In vivo estimation of succinate could help classifying a dubious lesion detected during surveillance and to demonstrate the causality of SDH deficiency in tumors identified in SDHx-mutated patients. This would be particularly helpful in cases for which surgery is not necessary, such as prolactin-secreting pituitary adenomas recently described in SDHx mutation carriers (37). Because of the small size of these tumors, the feasibility of SUCCESS may however be more difficult in these cases.
Animal experiments demonstrated that the area under the succinate peak of the 1H-MRS spectra is correlated with the concentrations of succinate measured in the resected tumors by GC-MS. Future studies in larger groups will be needed to show whether this correlation holds also in patients. If this turned out to be the case and given that succinate concentrations in tumors reflect the metabolic activity of SDH-deficient tumor cells, then SUCCES would produce a quantifiable surrogate marker of radiation and/or chemotherapy efficacy for the patients.
Interestingly, other metabolites have been shown in vitro to discriminate between different types of inherited PPGL (20). In future studies, it may also be addressed whether these metabolites can also be observed by 1H-MRS and used as supplemental tools. For example, Imperiale and colleagues recently reported in a case of sporadic pheochromocytoma that catecholamines are indeed detectable by 1H-MRS (38).
The present proof-of-concept study has shown that SUCCES is highly sensitive, reliable, and specific for the detection of the SDHx mutations that lead to inhibition of SDH activity. The next step to fully define the place of this new method in the clinical management of PPGL is to test the method in larger series of patients and define the best conditions for routine clinical applications. In that respect, we occasionally observed, in both groups of patients, a blunt signal centered at 1.2 ppm. This signal corresponds to adipose tissue surrounding the tumor that is not always straightforward to avoid even with strict intratumor positioning of the VOI in which 1H-MRS is performed. Because MRS data are usually displayed with a y scale normalized on the highest peak of the spectrum, the presence of a significant lipid signal may modify the threshold for succinate detection in small lesions. Averaging more spectra increases the signal-to-noise ratio but also increases scan duration, which may not be applicable to all patients. For abdominal tumors, respiratory gating should be considered to reduce the lipid peak and improve the quality of spectra, as previously reported for in vivo catecholamine detection (38). Using 512 scan averages appears to be sufficient for reliable succinate detection in tumors with SDHx genes mutation. However, this may limit the quality of spectra for small or highly necrotic tumors, as shown in the case of patient #3 (Supplementary Fig. S3B), for whom successful interpretation could only be achieved after 1,024 scan averages. Therefore, the minimal tumor size for reliable measurements of succinate needs to be addressed in future prospective studies. Fortunately, this 1H-MRS sequence is easy to implement in any clinical MRI scanner using standard hardware and software already in place in many imaging departments.
Finally, it is noteworthy that the succinate peak was always associated in human PPGL with a peak resonating at 3.2 ppm on the 1H-MRS spectra, most probably corresponding to choline. Such a peak was only seen in tumors from the patients carrying SDHx mutations. Previous in vitro NMR studies never reported such a choline increase in SDH-related tumors. However, a similar peak is also observed in the spectra of both SDHx-mutated paraganglioma evaluated by in vivo 1H-MRS in the Varoquaux and colleagues study (33). The accuracy of this observation will need to be further validated both in vivo and in vitro, in SDH as well as in other oncogenic mutations of metabolic pathways. It is worth noting that choline is a methyl donor in the S-adenosylmethionine pathway involved in DNA and histone methylation. Hence, if confirmed, the choline peak that we observed here may be related to the disrupted methylation phenotype recently identified in SDH-deficient tumors (13).
In conclusion, we present here a robust and simple method that can be used routinely to demonstrate the presence of succinate in the tumors of patients with PPGL. Considering its excellent sensitivity, specificity, and innocuousness, SUCCES deserves to be tested in large multicentric series to define its place in the clinical guidelines of PPGL management as well as in other SDH-related tumors such as GIST and renal clear cell carcinomas.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Conception and design: C. Lussey-Lepoutre, A. Morin, L. Amar, F. Zinzindohoué, A.-P. Gimenez-Roqueplo, J. Favier, B. Tavitian
Development of methodology: C. Lussey-Lepoutre, A. Bellucci, A. Morin, L. Amar, M. Janin, G. Autret, C.-A. Cuenod, J. Favier, B. Tavitian
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): C. Lussey-Lepoutre, A. Bellucci, A. Morin, A. Buffet, L. Amar, M. Janin, C. Ottolenghi, F. Zinzindohoué, G. Autret, N. Burnichon, E. Robidel, B. Banting, S. Fontaine, C.-A. Cuenod, P. Benit, P. Rustin, P. Halimi, J. Favier
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): C. Lussey-Lepoutre, A. Bellucci, N. Burnichon, P. Halimi, B. Tavitian, J. Favier
Writing, review, and/or revision of the manuscript: C. Lussey-Lepoutre, L. Amar, M. Janin, C. Ottolenghi, F. Zinzindohoué, G. Autret, N. Burnichon, C.-A. Cuenod, P. Halimi, L. Fournier, A.-P. Gimenez-Roqueplo, J. Favier, B. Tavitian
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): C. Lussey-Lepoutre, A. Morin, F. Zinzindohoué, E. Robidel, B. Banting, S. Fontaine, L. Fournier, A.-P. Gimenez-Roqueplo, J. Favier
Study supervision: C. Lussey-Lepoutre, A.-P. Gimenez-Roqueplo, J. Favier, B. Tavitian
The authors thank Prof. Pierre-François Plouin and Dr. Guillaume Bobrie for their clinical contribution to the study and Prof. Catherine Oppenheim and Stephanie Lion for sharing their expertise. They also thank Daniel Balvay for helpful discussions, Brigitte Lambert (Radiology, HEGP) and Marion Uettwiller (General Electrics Healthcare) for technical assistance, and Catherine Tritscher for administrative assistance. They thank the technical staff of the Genetic department of HEGP (especially Françoise Le Quellec, Caroline Travers, Nirubiah Thurairajasingam) led by Prof. Xavier Jeunemaitre, Jean-Michaël Mazzella, and Samir Jocelyn Do Rego for their contribution to the study. They thank Daniel Tennant for helpful discussion.
This work received funding from the Cancer Research for Personalized Medicine—CARPEM project (Site de Recherche Intégré sur le Cancer- SIRIC), the Agence Nationale de la Recherche (ANR-2011-JCJC-00701 MODEOMAPP), the European Union Seventh Framework Programme (FP7/2007-2013) under grant agreement no. 259735, and the Institut National du Cancer et la Direction Générale de l'Offre de Soins (INCa-DGOS_8663) for the COMETE network. C. Lepoutre-Lussey is funded by the CARPEM project. A. Bellucci received a fellowship from Fondation pour la Recherche Médicale. A. Buffet received a fellowship from the ITMO Cancer - Plan Cancer 2014-2019.