Identification of Succinate Dehydrogenase Gene Variant Carriers by Blood Biomarkers

Abstract Background Carriers of germline pathogenic variants (PVs) in succinate dehydrogenase genes (SDHx) are at risk of developing tumors, including paragangliomas, gastrointestinal stromal tumors, and renal cell carcinomas. Early tumor detection is paramount for improved clinical outcome. Blood-based biomarkers could aid in identifying individuals with PVs early and provide functional evidence in patients with variants of unknown significance. Methods Blood plasma, urine, peripheral blood mononuclear cells, and erythrocytes from patients with and without SDHx PVs were investigated for central carbon metabolites. These were measured by liquid chromatography–tandem mass spectrometry and nuclear magnetic resonance spectroscopy and included among others, succinate, fumarate, α-ketoglutarate, and lactate. Results Plasma succinate to fumarate ratios effectively distinguished tumor-bearing and asymptomatic patients with and without SDHx PV with promising diagnostic performance (areas under the receiver operating characteristic curve 0.86-0.95), although higher levels were noted in individuals with SDHB PV. Metabolites in urine and in peripheral blood mononuclear cell extracts were largely similar between groups. Erythrocytes showed strong metabolic alterations in patients with SDHx PV compared to controls, with 8 of 13 low-molecular organic acids being significantly different (P < .05). The lactate-α-ketoglutarate-ratio of erythrocytes identified individuals with SDHx PV equally well as plasma, with a sensitivity and specificity of 92% (AUC 0.97). Conclusion Blood biomarkers have been underutilized for identifying carriers of SDHx PV or to validate variants of unknown significance. Our findings advocate for further investigation into a combined approach involving plasma and erythrocytes for future diagnostic strategies.

Succinate dehydrogenase (SDH) is an enzyme complex located at the inner mitochondrial membrane, where it participates in metabolism of the Krebs cycle and also facilitates electron transfer within the respiratory chain as complex II.SDH comprises 4 subunits: SDHA is catalytically active and converts succinate to fumarate.SDHB contains Fe-S clusters for electron transfer, and SDHC and SDHD are membrane anchors.Heterozygous loss-of-function germline variants in succinate dehydrogenase genes (SDHA, SDHB, SDHC, SDHD, and assembly factor SDHAF2) are responsible for a number of tumor diseases, foremostly paragangliomas of the sympathetic or parasympathetic nervous system and pheochromocytomas (together referred to as PPGL).Carriers of germline pathogenic variants (PVs) in SDHx can also develop gastrointestinal stromal tumors, renal cell carcinomas, and rarely, pituitary adenomas or thyroid cancer [1].Due to this risk profile, lifelong tumor screening is recommended for SDHx PV carriers [2].Potential benefits of such a screening approach are evident from the literature.Early surveillance resulted in reduced metastatic spread in patients with SDHx PVs and in increased metastasis-free and overall survival in patients with PVs in SDHB [3,4].
The basis of any surveillance program requires correct identification of individuals at risk.Low penetrance of disease can hinder the recognition of family transmission leading to SDHx PV detection earliest at the stage of tumor presentation [5][6][7][8].Furthermore, variant classification is especially difficult for missense variants in SDHx genes.Currently 3334 variants of uncertain significance (VUS) or with conflicting interpretations are listed for SDHx genes and the assembly factor SDHAF2 in the ClinVar database (accessed March 8, 2024).
Several methods have been described to identify patients with tumors due to SDHx PVs, including ex vivo measurements of the succinate to fumarate ratio (SFR) in tumor tissue or the SDHB protein by immunohistochemistry as well as in situ detection of succinate by proton magnetic resonance spectroscopy [9].As complete loss of SDH activity results in a blockade of the Krebs cycle with strong accumulation of the precursor succinate and a decrease of the product fumarate, elevated succinate or SFR levels in the tumor are important diagnostic parameters that can aid variant classification [10,11].More recently, it was demonstrated that serum succinate levels are elevated in patients with SDHB PVs and PPGL as well as in asymptomatic carriers [12,13].Our study aims to validate and extend these results by not only looking at the liquid parts of blood, that is, the serum or plasma, but also at cellular components, such as peripheral mononuclear cells (PBMCs) and erythrocytes.Therefore, we investigated these parameters in patients with a history of PPGL before and/or after tumor removal.

Study Participants, Clinical Data, and Sample Collection
Patients with a past history or presence of PPGL were enrolled in 3 clinical studies, the multicenter prospective monoamineproducing tumor study (PMT), PROSPHEO (NCT03344016), or through registry and biobanking protocols of the European Network for the Study of Adrenal Tumours (ENSAT).Two separate patient cohorts were investigated (Fig. 1).Cohort A contained patients enrolled under PMT, from whom heparin plasma (n = 38) and 24-hour urine specimens (n = 36) were collected before and/or after the removal of PPGLs (numbers are detailed in Fig. 1).Cohort B consisted of patients enrolled in Dresden through ENSAT and PROSPHEO, from whom EDTA plasma (n = 45) and blood cells (n = 25) were collected before and/or after the removal of PPGLs (numbers are detailed in Fig. 1 and Supplementary Tables S1-S7 [14]) or from asymptomatic SDHx PV carriers (n = 4).Only for 5 patients of cohort B were paired samples of before and after surgery available.Blood samples were taken after an overnight fast.Clinical data included sex, age, germline PV, body mass index (BMI), presence of diabetes mellitus, and tumor volume (determined according to at least 2 recorded dimensions [15]).This work was conducted in accordance with the principles of the Declaration of Helsinki.All patients were enrolled into clinical protocols approved by local review boards in Dresden, Munich, Würzburg, Warsaw, and Nijmegen.All patients signed informed consent.

Genetic Testing
Germline PVs in SDHA, SDHB, SDHC, SDHD, SDHAF2, and other genes predisposing to PPGL (VHL, TMEM127, MAX, NF1, and RET) were identified using next-generation sequencing and/or Sanger sequencing and multiplex ligationdependent probe amplification or custom array comparative genomic hybridization detecting large-scale deletions.Genetic testing was performed in all patients either at local centers or at the CNIO, Madrid, Spain.Classification of germline variants aligned with current practice guidelines [16].Patients included in this study had variants in SDHx genes classified as pathogenic or likely pathogenic (Supplementary Tables 1-7 [14]).For 2 variants of uncertain significance in SDHA, additional functional data, namely, measurements of SFR in tumor tissue from included patients, were available and led to reclassification as likely pathogenic, since SFRs were strongly elevated compared to a previously determined cutoff [17].For 2 patients of cohort A the specific gene variant was unknown.

Preparation of Plasma and Urine Samples
EDTA and heparin blood samples (10 mL) were centrifuged at 4200g for 15 minutes at 4 °C (Heraeus Multifuge 1S-R).Plasma and 24-hour urine were aliquoted and stored until further analysis at −80 °C.
Samples of 20 µL undiluted plasma and 20 µL of a 1:10 dilution in water were homogenized with 10 µL of internal standard solution [18] and 400 µL of ice-cold methanol to facilitate protein precipitation.After centrifugation (16 000g, 5 minutes, 4 °C), clear supernatants were transferred into a 96-well plate and dried using a vacuum-assisted centrifuge (35 °C, 3 hours), followed by reconstitution in 100 µL of initial mobile phase (0.1% formic acid in water/0.1% formic acid in acetonitrile, 99/1, v/v) and sample injection into the liquid chromatography-tandem mass spectrometry (LC-MS/MS) system.

Preparation of Erythrocytes and Metabolite Extraction
EDTA blood samples were centrifuged at 4200g for 15 minutes at 4 °C (Heraeus Multifuge 1S-R) within 4 hours of sample collection and refrigerated storage.Erythrocytes were then collected with care from the very bottom of the tube to avoid contamination with leukocytes.Purity of the erythrocyte preparation was checked visually under a microscope.A volume of 5 µL erythrocytes was mixed with 500 µL of ice-cold methanol and stored at −20 °C for 20 minutes.After centrifugation at 16 000g for 5 minutes at 4 °C, the clear supernatant was divided into 2 aliquots (20 µL and 450 µL), transferred to new reaction tubes, and mixed with 10 µL of internal standard solution.Samples were dried in a vacuum centrifuge (Savant Speedvac SPD 210) and stored at −80 °C until further processing.Additionally, 50 µL of erythrocytes were stored in 1 mL freezing medium (PromoCell) at −80 °C for protein isolation.

Preparation of Peripheral Blood Mononuclear Cells
Lymphoprep™ density gradient medium (Stem Cell Technologies) was aliquoted at 15 mL in a 50-mL tube.EDTA blood was diluted 1:2 with phosphate-buffered saline (PBS, Gibco).Subsequently, the dilution was layered onto the Lymphoprep density gradient medium.The tube was centrifuged at 800g for 20 minutes with disabled deceleration (Rotina 380, Hettich centrifuges).PBMCs located on top of the density gradient medium were collected and transferred to a new tube.PBMCs were washed twice with 15 mL PBS and centrifuged at 500g for 10 minutes.After the second wash step, cells were resuspended in 5 mL PBS and counted using a Neubauer chamber (C-Chip, NanoEntek).The cell suspension was centrifuged for another 10 minutes at 120g with decreased acceleration speed to remove thrombocytes.PBMCs (1 × 10 6 cells/mL) were lysed in 500 µL of ice-cold methanol, kept at −20 °C for 20 minutes and centrifuged at 16 000g for 5 minutes at 4 °C.The clear supernatant was divided in 2 aliquots (20 µL and 450 µL), transferred to new reaction tubes, and mixed with 10 µL of internal standard solution.Samples were dried in a vacuum centrifuge and stored at −80 °C until further processing.

Protein Isolation From Erythrocytes and Western Blot
Preparation of the membrane fraction Erythrocytes were thawed, immediately transferred to cell culture medium (RPMI 1640, Gibco) containing 10% horse serum, 5% fetal bovine serum, and 1% Glutamax and centrifuged at 4200g for 10 minutes.The supernatant was removed and the pellet was washed in cold PBS.Following the protocol by Wilson et al [19], cells were washed twice with iso-osmotic buffer (103 mM disodium phosphate [Na 2 HPO 4 ], 155 mM monosodium phosphate [NaH 2 PO 4 ], pH 7.4) and afterwards lysed in 10 mL cell lysis buffer (1:20 dilution of iso-osmotic buffer, 1x protease inhibitor cocktail (Sigma Aldrich), 0.5 mM phenylmethylsulfonylfluoride).Cells were centrifuged at 4200g, 4 °C for 10 minutes.This lysis step was repeated until the red cell pellet turned completely white, resembling the membrane fraction of erythrocytes.

Sample preparation for SDS-PAGE
The membrane fraction was resuspended in 35 µL radioimmunoprecipitation assay buffer (RIPA, Serva) containing 1x protease inhibitor cocktail and chilled on ice for 20 minutes.After centrifugation at 16 000g, 5 minutes, 4 °C, supernatant was transferred to a new reaction tube, mixed with 4-fold Laemmli buffer (Bio-Rad) and 5% β-mercaptoethanol and denatured at 100 °C for 5 minutes.

Nuclear Magnetic Resonance Spectroscopy
Frozen samples were thawed at room temperature for 30 minutes before being mixed with phosphate buffer containing internal reference trimethylsilylpropanoic acid (300 µL each) and urine samples with a phosphate buffer (540 µL to 60 µL) to a volume of 600 µL.The resulting suspension was transferred into nuclear magnetic resonance (NMR) sample tubes and loaded onto the machine.All NMR measurements were run on a Bruker 600 MHz Avance III Neo equipped with a BBI Probe and a Bruker SampleJet robot with a cooling system for sample storage at 4 °C.The plasma samples were measured at 310 K and urine samples at 300 K.A full quantitative calibration was completed before the measurement.All measurements followed the Bruker in vitro diagnostics (IVDr) standard operating procedures and No Calculations based on measurements of at least 2 dimensions as described elsewhere (15).

Statistics and Visualization
Data were analyzed and visualized with the software package JMP Pro 17

Study Population
This study included 2 cohorts of patients, who either had PPGL themselves or had family members with PPGL and hence received genetic testing (Fig. 1, Supplementary Tables 1-7 [14]).In cohort A, all patients had PPGL, 15 individuals had heterozygous germline PVs in SDHx genes, 2 patients had somatic SDHx aberrations, and 21 patients had no PVs in SDHx.Cohort B consisted of 20 individuals with germline PVs in SDHx, from whom 4 were asymptomatic, 1 patient with SDHC promoter methylation [21], and 24 patients without PVs in SDHx, who had a history of PPGL.Distribution of sexes was similar in both cohorts with a roughly equal split between male and female in the SDHx PV group, and an overrepresentation of female individuals in the group of patients without SDHx PVs (Table 1).Sampling age was lower for patients with SDHx PVs in both cohorts.All clinical parameters referred to in this study, including tumor volume, BMI, presence of diabetes mellitus type 2, and erythrocyte characteristics, are summarized in Table 1.

Plasma SFR Best Separates Patients With and Without SDHx PV
Metabolites were measured in heparin plasma (cohort A) using 2 technologies: a LC-MS/MS approach for the assessment of low-molecular organic acids of the Krebs cycle, and NMR spectroscopy, which quantifies a broader spectrum of metabolites.All raw data are presented in Supplementary Tables 1-7 [14].
The SFR determined by LC-MS/MS was significantly different between patients with and without SDHx PV, both in plasma collected preoperatively and postoperatively (Fig. 2A and 2B).
The SFR was generally lower after tumor removal in 11 of 15 patients with germline SDHx PV and in 1 of 2 patients with somatic SDHx aberration (Fig. 2C).Metastatic disease was not the reason for persistently high or higher than baseline SFRs.A failed drop in SFR occurred more often in patients with SDHD PV, but not exclusively (Fig. 2D).Furthermore, decreased SFR after surgery occurred in 8 of 9 female patients, but only in 3 of 6 male patients.Preoperative SFR correlated significantly with tumor volume, indicating that at least some of the plasma metabolites originate from tumor cells (Fig. 2E).Significantly higher SFRs were measured preoperatively in plasma from patients with SDHA or SDHB PV compared to patients with SDHC or SDHD PV (Fig. 2F).Succinate and metabolite ratios containing succinate were the only significant differences between patients with and without SDHx PV preoperatively and postoperatively (Table 2).NMR spectroscopy was unable to detect succinate or fumarate in plasma.In postoperative samples, NMR spectroscopy identified glutamine as the only metabolite to be significantly different between patients with and without SDHx PV (adjusted P = .026).Glutamine separated these groups with an area under the ROC curve (AUC) of 0.849.

Diagnostic Performance of SFR in Plasma
Analogous to investigations in heparin plasma, metabolites were measured in EDTA plasma (cohort B) by LC-MS/MS.The SFR separated patients with and without SDHx PV preoperatively and postoperatively with high statistical significance (Fig. 3A and 3B).Patients with SDHB PV showed the highest SFRs (Fig. 3C).Type of gene was the only significantly associated variable by regression analysis among the parameters, tumor presence, sex, age, diagnosed diabetes mellitus, and BMI.In both matrixes the SFR showed high diagnostic performance with AUCs of 0.867 to 0.947 (Table 2).Other possible ratios are succinate/citrate and succinate/malate with AUCs of 0.794 to 0.938 and 0.775 to 0.956, respectively.Succinate alone had AUCs of 0.699 to 0.907.Since our cohorts only included one patient with germline FH PV, an analysis of metabolic parameters for the identification of such individuals is not possible; however, asparagine, asparagine/2-hydroxyglutarate, fumarate/ 2-hydroxyglutarate, malate/2-hydroxyglutarate, and glutamine/ 2-hydroxyglutarate are possible markers that should be investigated in a larger cohort (Table 3).

Urinary Metabolites Do Not Discriminate Between Patients With and Without SDHx PV
Given that urinary succinate levels were more than 10 times higher than those in plasma (median 62 [range, 11-347] Table 2. Diagnostic power of succinate and ratios to differentiate between patients with and without SDHx PV in the presence (pre-OP) and absence (post-OP) of PPGL mmol/L vs 4 [2-52] mmol/L), NMR spectroscopy also detected this metabolite.Levels of succinate and the SFR were in a similar range in patients with and without SDHx PV.No significant differences between patients with and without SDHx PV were detected in urine.

Erythrocytes but Not PBMCs Differ in Their Metabolism Between Patients With and Without SDHx PV
Succinate or the SFR did not show differences between patients with and without SDHx PV in PBMCs (data not shown) or erythrocytes (Fig. 4A).No significant metabolic alterations were measured in PBMCs, whereas strong differences were detected in erythrocytes of patients with compared to those without germline SDHx PV (Fig. 4B-4I).Of 13 quantified metabolites, 8 were significantly different between groups, including fumarate, malate, isocitrate, cis-aconitate, α-ketoglutarate, 2-hydroxyglutarate, lactate, and glutamine.Lactate/ α-ketoglutarate and lactate/2-hydroxyglutarate showed the best diagnostic separation of patients with and without SDHx PV with an AUC of 0.968 (Table 4, Fig. 5A).Interestingly, these erythrocyte parameters appear to be to a lesser extent influenced by gene type (Fig. 5B).Combination of SFR in EDTA plasma and lactate/α-ketoglutarate in erythrocytes yielded a marginal improvement of specificity over either parameter alone, reaching a specificity of 100% and sensitivity of 92% (Table 5).Analysis of patient blood counts did not yield any significant differences in erythrocyte parameters between carriers with SDHx PVs and other patients (Table 1).Furthermore, we did not detect any consistent differences in SDHB protein levels in membrane fractions of erythrocytes between individuals with and without SDHx PV (Fig. 6).

Discussion
This study shows for the first time that not only plasma from patients with and without SDHx PVs is distinct, but also erythrocytes have different features.PBMCs, on the other hand, were largely unchanged.Our findings open up new diagnostic options for the identification of carriers with SDHx PVs and could aid in classification of VUS in SDHx genes.ClinVar currently has more than 3000 entries for VUS in SDHx genes (https://www.ncbi.nlm.nih.gov/clinvar,visited June 25, 2024).Relatives of patients with SDHx-related PPGL in whom a VUS was identified are not eligible for genetic testing and will not be screened regularly for tumors.In these cases, a bloodbased test could provide more clarity.Additionally, the proposed assay would be useful in settings where genetic testing is not available, or delayed, or patients refuse genetic analysis.Due to higher sensitivity of LC-MS/MS compared to NMR, LC-MS/MS is the preferred technique for measurements of carbon metabolites, especially in plasma, where succinate levels are 10-fold lower compared to urine.
Our investigation extends results from 2 previous studies that showed succinate or the SFR in serum is elevated in patients with SDHB PVs, both in the presence of PPGL and in asymptomatic patients [12,13].Patients with PVs in other subunits of succinate dehydrogenase did not show a significant increase in those studies.Our results confirm that individuals with SDHB but also SDHA PVs have the highest plasma succinate levels.This influence was even stronger than the presence of PPGL.Whether plasma SFR will be a useful biomarker for tumor development in carriers of SDHx PVs and whether this assay could reduce the need for expensive wholebody imaging will have to be investigated in the future.
Other previously proposed parameters influencing succinate levels, such as obesity and diabetes mellitus type 2, were not significant factors in our study [22].The source of succinate in tumor-free SDHx PV carriers is unknown but might originate from metabolically active and energy-demanding cells, such as liver or muscle.Recently, succinate was described as a paracrine regulator of muscle remodeling after exercise [23], raising the question whether carriers of SDHx PVs have altered responses to exercise.Lower levels of plasma succinate in individuals with SDHC and SDHD PVs could indicate that a lesser expression of the membrane bound subunits SDHC and SDHD might be more easily compensated than reduction in subunits facilitating catalytic activity.
Correlation of plasma SFR with tumor volume supports a previous study, in which serum succinate was shown to function as a biomarker for tumor recurrence and tumor burden in 3 patients with metastatic SDHB-related PPGL [12].SFR could be easily measured in the same sample and with the same instrumentation used for plasma metanephrines [24] and would provide another parameter for evaluating treatment success or recurrence in patients with SDHx-related PPGL.Our analysis indicates that, beyond SFR, also other ratios with    succinate could be utilized, specifically succinate/citrate and succinate/malate, which showed similar results.
Despite succinate elevations in plasma, urinary succinate or SFR were similar between patients with and without SDHx PVs.Urinary sources of succinate might originate from cell populations less responsive to SDH loss and might only occur in pathological settings.To this end, increases in urinary succinate were linked previously to reduced mitochondrial metabolism and more severe kidney disease [25].
A previous study reported on increased succinate and lactate in neutrophils of patients with SDHB PVs [26].PBMCs, on the other hand, did not show any difference between patients with and without SDHx PVs in our study.As succinate drives proinflammatory responses in macrophages [27], there may be regulatory safety nets to prevent succinate elevations at baseline.Granulocytes were not investigated in the present study.
Although mature erythrocytes are devoid of mitochondria, SDHB is present in their membrane fraction [19].Furthermore, it was demonstrated that high altitude affects levels of carboxylic acids of the central carbon metabolism in erythrocytes, including α-ketoglutarate, succinate, and others [28].Our study found that erythrocytes from individuals with SDHx PV have lower levels of fumarate, malate, isocitrate, cis-aconitate, α-ketoglutarate, 2-hydroxyglutarate, and glutamine, but higher levels of lactate compared to controls.Based on the increased lactate to α-ketoglutarate ratio, we hypothesize that glycolytic flux may be increased in erythrocytes from patients with SDHx PV by favoring glycolysis over the pentose phosphate pathway, which is responsible for maintaining antioxidant capacity [29].Mice with heterozygous deletions of different SDHx genes show an increased healthy life span under hypobaric oxygen with reduced red cell distribution width [30].This may be facilitated by an improved glycolytic flux that most likely results in higher energy levels under low oxygen conditions in erythrocytes with SDHx loss compared to controls.A recent observation of an inheritance ratio of 0.6 for SDHB PVs suggests that heterozygous SDHB loss has developmental advantages [31].As different regions of the embryo are exposed to hypoxia during development [32], it could be conceivable that more hypoxia-resistant cells may be selected for.
The present study is a proof-of-concept for the combination of plasma and erythrocyte parameters, and it suffers from a number of limitations.Foremost, the relatively small number of patients included in our analysis complicates investigations toward factors with more subtle influence on plasma SFR levels.Furthermore, the lack of a large comparison group of healthy individuals hinders the establishment of reference intervals for diagnostic use.The present study only analyzed blood samples from a maximum of 2 time points and did not follow individual patients through their course of disease.Hence, we cannot comment on how plasma SFR behaves in longitudinal follow-up, for which imaging information as comparison would be crucial.The latter were not included in the present analysis; instead, tumor burden was assessed based on volume calculations [15].Systematic investigations on what factors influence physiological levels of succinate and fumarate were not part of this study.To account for food-related effects, samples were taken in line with preanalytics for metanephrines after an overnight fast.As exercise and BMI were previously implicated in plasma succinate variations [22,23], individual levels of succinate may be associated with muscle mass and fat depots.
This study shows for the first time that not only liquid parts of a blood sample carry valuable information about potentially pathogenic variants in one of the SDHx genes, but also  erythrocytes have diagnostic power.As only a small number of patients (n = 25) were investigated for changes in erythrocyte metabolism, it will be important to follow up with a prospective study, as the described method could be valuable for early identification of PV carriers in all SDHx genes.Additionally, our results highlight that heterozygous PVs in SDHx genes lead to physiological changes due to loss of one SDHx allele.Further insights into these physiological adaptations might uncover whether carriers are at a higher or possibly lower risk for developing other more common diseases.

Figure 1 .
Figure 1.Overview of patient (pt.) cohorts and available material for LC-MS/MS and NMR.A detailed breakdown of patients can be found in Supplementary Tables 1-7 [14].

Figure 2 .
Figure 2. Succinate to fumarate ratio determined in heparin plasma by LC-MS/MS.Plasma was collected from patients with PPGL (cohort A) either preoperatively (Pre-OP, A) or postoperatively (Post-OP, B).C) Paired analysis between pre-OP and post-OP samples; P value represents Wilcoxon signed-rank test (1-tailed for SDHx, 2-tailed for non-SDHx).D) Decrease in the succinate to fumarate ratio according to genes in % (Kruskal-Wallis test).E) Correlation between pre-OP succinate to fumarate ratios and tumor volume (natural logarithm of value in cm 3 displayed).F. Succinate to fumarate ratio according to the affected gene comparing patients with germline SDHA (black dots) and SDHB (gray dots) PVs to patients with SDHC (black dots) and SDHD (gray dots) PVs (Kruskal-Wallis test).

Figure 3 .
Figure 3. Succinate to fumarate ratio determined in EDTA plasma by LC-MS/MS.Plasma was collected from patients of cohort B either preoperatively (Pre-OP, n = 18, A) or in patients without tumor burden (Post-OP, n = 31, removal of a high succinate outlier in the group of patients with SDHx PV, B).Tumor burden refers to the presence of primary tumors or metastases and was determined by anatomical and/or functional imaging.C) Succinate-fumarate-ratios are displayed according to the affected gene.Triangles depict patients with a PPGL, circles show patients without tumor burden.Patients with SDHx germline PV, SDHx somatic PV, and without SDHx PV are shown in red, orange, and blue, respectively.

a
Median and range in parentheses.

Figure 4 .
Figure 4. Krebs cycle metabolites in erythrocytes from patients with and without SDHx PVs.A-I) Erythrocytes were collected from patients (cohort B) in whom PPGL was present (triangles, light red) or absent (circles, dark red).Patients with SDHx germline PV and without SDHx PV are shown in red and blue, respectively.

Figure 5 .
Figure 5.The lactate/α-ketoglutarate-ratio from erythrocytes separates patients with SDHx PVs.A) Erythrocytes were collected from patients (cohort B) in whom PPGL was present (triangles, light red) or absent (circles, dark red).Patients with SDHx germline PV and without SDHx PV are shown in red and blue, respectively.Gray dashed lines indicate cutoff according to Youden index (186).As ratios of lactate/ketoglutarate and lactate/2-hydroxyglutarate show similar results, only the first one is displayed here.B) Lactate/α-ketoglutarate does not significantly differ between patients with different SDHx genes.Since the lactate/2-hydroxyglutarate ratio shows exactly the same results as lactate/α-ketoglutarate, only the latter is displayed in this figure.

Table 4 . Most significantly different metabolite ratios in erythrocytes from patients with vs without SDHx PVs Parameter a P value b AUC of ROC curve c
Abbreviations: 2HG, 2-hydroxyglutarate; AUC, area under the curve, PV, pathogenic variant; ROC, receiver operator characteristic.aOnly most significant metabolite ratios with AUC > 0.9 are listed.bLikelihood ratio test.c 95% CIs in parentheses.

Table 5 . Diagnostic performance of metabolite ratios in plasma and erythrocytes to identify patients with SDHx PVs Parameter AUC of ROC curve Specificity Sensitivity
Specificity/sensitivity were calculated based on the Youden index of ROC curves.Abbreviations: AUC, area under the curve, PV, pathogenic variant; ROC, receiver operator characteristic.