Abstract
Mutations in genes encoding enzymes in the tricarboxylic acid cycle (TCA, also known as the Krebs cycle) have been implicated as causative genetic lesions in a number of human cancers, including renal cell cancers, glioblastomas and pheochromocytomas. In recent studies, missense mutations in the succinate dehydrogenase (SDH) complex have also been proposed to cause differentiated thyroid cancer. In order to gain mechanistic insight into this process, we generated mice lacking the SDH subunit D (Sdhd) in the thyroid. We report that these mice develop enlarged thyroid glands with follicle hypercellularity and increased proliferation. In vitro, human thyroid cell lines with knockdown of SDHD exhibit an enhanced migratory capability, despite no change in proliferative capacity. Interestingly, these cells acquire stem-like features which are also observed in the mouse tumors. The stem-like characteristics are reversed by α-ketoglutarate, suggesting that SDH-associated tumorigenesis results from dedifferentiation driven by an imbalance in cellular metabolites of the TCA cycle. The results of this study reveal a metabolic vulnerability for potential future treatment of SDH-associated neoplasia.
Introduction
Thyroid cancer is the most common endocrine malignancy with a rapidly increasing incidence rate as well as a high degree of familiality for certain subtypes (Leenhardt et al. 2004, Davies et al. 2015). Thyroid cancer can be divided into the rare (5% of the cases) medullary thyroid cancer (MTC), arising from parafollicular C-cells, and the more common (95%) epithelial thyroid cancer (also called non-medullary thyroid cancer, NMTC), arising from the thyrocytes themselves (Kitahara & Sosa 2016). The majority of patients with NMTC exhibit well-differentiated pathologies of either papillary (PTC) or follicular (FTC) subtype, and almost all the increase in the incidence is observed in PTC (Pringle et al. 2014). The majority of PTC and FTC are sporadic, while the familial component is believed to be 5% of the cases (Morrison & Atkinson 2009, Nose 2011). Familial cases may occur as part of rare high penetrance genetic syndromes or from more prevalent low-penetrance susceptibility genes (Nagy et al. 2004, Tomsic et al. 2015). Cowden syndrome (CS) is an autosomal dominant disorder that is characterized by high risk of breast cancer, thyroid neoplasms, uterine tumors, gastrointestinal hamartomas and less frequent tumors in other solid organs (Marsh et al. 1998, Hobert & Eng 2009). Most cases of CS and a small number of Cowden syndrome-like (CSL) cases are associated with inactivating mutations in the Phosphatase and tensin homolog (PTEN) tumor suppressor gene.
Recently, mutations in genes encoding the succinate dehydrogenase (SDH) subunits were identified as additional susceptibility genes for CS, comprising about 10% of CS and CSL phenotypes (Ni et al. 2008, Yu et al. 2015). SDH is a heterotetrameric nuclear-encoded mitochondrial protein which comprises four subunits encoded by the four autosomal genes SDHA, SDHB, SDHC and SDHD. These genes are known tumor suppressors owing to their established connection to the syndrome of inherited pheochromocytoma (PHEO) and paraganglioma (PGL) (Bardella et al. 2011, Letouze et al. 2013). Genetic variants in SDHB and SDHD have been identified in a subset of CS/CSL patients, and confer high risk of breast, thyroid and other cancers. Furthermore, downregulation of SDH subunits has been observed in both PTC and FTC and has been shown to correlate with poorer prognosis. Interestingly, CS/CSL patients with variants in SDHD exhibit an increased risk for PTC, while individuals with mutations in PTEN have a predilection to both PTC and FTC (Ni et al. 2015).
SDH is a component of the tricarboxylic acid cycle (TCA cycle, also known as the citric acid or Krebs cycle), as it oxidizes succinate to fumarate and leads to electron transport to ubiquinone in the electron transport chain (ETC). SDH was the first mitochondrial enzyme that proved that mutations in the TCA cycle genes can in fact act as drivers of tumorigenesis. Later, mutations leading to dysfunction of other TCA cycle proteins such as SDH assembly factor 2 (SDHAF2), Fumarate hydratase (FH) and Isocitrate dehydrogenase (IDH) were found to be associated with multiple types of cancer formation (Cantor & Sabatini 2012). The spectrum of tumors caused by mutations in each of these genes differs, although there may be some overlap among the syndromes. Inherited or somatic mutations in any of the four subunits of SDH can lead to PHEO, PGL, renal cell carcinoma, gastrointestinal stromal tumor, thyroid cancer and breast cancer (Bardella et al. 2011, Xiao et al. 2012, Letouze et al. 2013, Millan-Ucles et al. 2014, Williamson et al. 2015).
Dysfunction of these metabolic enzymes has provided new insight into the observation that the majority, although not all cancers, show a preference toward anaerobic glycolysis over oxidative phosphorylation. This phenomenon, known as the Warburg effect, enables cells to grow under hypoxic conditions (Archetti 2015). This metabolic alteration was initially thought to be an adaptive mechanism to overcome the hypoxic conditions in tumors, but newer data suggest that metabolic dysfunction is itself a driver of tumorigenesis. Mechanisms that have been proposed to explain how loss of SDHD leads to thyroid tumorigenesis include (1) an increase in the production of reactive oxygen species (ROS), (2) activation of a hypoxia-like pathway under normoxic conditions (pseudohypoxia) and (3) genetic and epigenetic alterations due to the presence of oncometabolites (MacKenzie et al. 2007, Salminen et al. 2014, Jardim-Messeder et al. 2015, Yu et al. 2015). Specifically, several lines of genetic and biochemical evidence suggest that citrate, succinate, fumarate and α-ketoglutarate (α-KG, also known as 2-oxoglutaric acid or 2-OG) may serve as oncometabolites to promote tumorigenesis by inhibiting a family of epigenetic modulator enzymes (Letouze et al. 2013, Carey et al. 2015, Sciacovelli & Frezza 2016).
To evaluate the contribution of SDHD loss on thyroid tumorigenesis, we generated tissue-specific knockout of Sdhd in the mouse thyroid gland. These in vivo studies were complemented by in vitro analyses of human thyroid cancer cells with knockdown of SDHD. Together, these studies reveal the ability of SDHD/Sdhd loss in promoting thyroid neoplasia and provide new mechanistic insights for these observations.
Methods
Animal strains, husbandry and maintenance
The use of animals was in compliance with federal and Ohio State University Laboratory Animal Resources regulations. SdhdloxP/loxP mice (Millan-Ucles et al. 2014) were crossed with Thyroid Peroxidase (Tpo)-cre (Kusakabe et al. 2004) to generate Sdhd-TpoKO. In addition, PtenloxP/loxP mice (Pten-TpoKO) were generated as described previously (Pringle et al. 2012). Tpo-cre, SdhdloxP/loxP and PtenloxP/loxP were mated to generate Tpo-Cre; SdhdloxP/loxP PtenloxP/loxP (SP-TpoKO) mice. The experiments were performed using littermate mice from a mixed C57BL/6 and FVB genetic background.
Ultrasonography
All mice were imaged with a VisualSonics Vevo 2100 (VisualSonics, Toronto, CA, USA) every 3 months up to 1 year. MS550D transducer (22–55 MHz) was used with 3D-mode imaging to determine the size of the thyroid. The volume of both thyroid lobes was calculated using Vevolab 2.1 software in a blinded fashion.
Follicular area measurement
The hematoxylin and eosin (H&E)-stained sections were imaged at 20× magnification. The area of follicles was determined by measuring the luminal surface using the ImageJ software as described previously (Yeager et al. 2007).
Cell culture and reagents
FTC133 (human follicular thyroid carcinoma) cells were maintained in DMEM media with 10% FBS and penicillin/streptomycin. NthyOri 3.1 (human follicular epithelial thyroid cells) were maintained in RPMI media with 10% FBS and penicillin/streptomycin. Octyl-α-ketoglutarate was purchased from Cayman Chemicals. Lentiviral-based ShRNA (sequences available on request) for SDHD knockdown was obtained from Sigma Aldrich. Lentiviral supernatant was produced from phoenix cells transfected with Fugene (Promega), packaging mix (Sigma) and viral vector. Stable SDHD-knockdown thyroid cells were generated by puromycin selection after transducing with ShRNA lentivirus.
Primary thyrocyte isolation
Primary thyroid cells were isolated from mouse tumors by enzymatic dissociation using Collagenase/Hyaluronidase solution, Dispase and DNAse using protocols adapted from manufacturer (StemCell Technologies, Vancouver, Canada). Briefly, the thyroid was dissected and chopped using automatic tissue chopper and placed in Collagenase/Hyaluronidase solution for 3 h at 37°C with agitation. The sample was centrifuged at 350g for 5 min and the pellet was resuspended in 0.25% trypsin-EDTA on ice for 1 h. After centrifugation, the cell pellet was resuspended in ice-cold HBSS with 2% serum. After centrifugation, cells were dissociated with pre-warmed Dispase and Dnase I, resuspended ice-cold HBSS with 2% serum and filtered through a 40-µM cell strainer. Cells were kept on ice until use.
ALDEFLUOR assay
The ALDEFLUOR kit (StemCell Technologies) was used to isolate the cell population with a high aldehyde dehydrogenase (ALDH) activity. Cells were obtained from cultured cells or mouse thyroid tumors. Optimal conditions for the experiment were determined by titrating cell number and incubation time. Cultured cells and primary tumor cells were resuspended in ALDEFLUOR assay buffer containing ALDH substrate (BAAA, 5 µL per 1 × 105 cells/mL) and incubated for 20 min at 37°C. An aliquot of each sample was treated with diethylaminobenzaldehyde (DEAB, 5 µL per 1 × 105 cells/mL), a specific ALDH inhibitor as a negative control. Propidium iodide was used for cell viability. Cells were kept on ice after the ALDEFLUOR reaction was completed until analyzed on BD LSR II flow cytometer. DEAB-treated negative control sample was used as a gating control and gate was set to include 0.5% ALDHbright cells on DEAB treatment for each sample.
Western blotting
The proteins were run on SDS-PAGE gel by a standard procedure, and the membranes were probed with the indicated antibodies: Millipore, SDHD (ABT110); Cell Signaling Technology, GAPDH (2118). Signals were detected with SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Fisher).
Mass spectrometry
Organic acid and phosphorylated compounds were extracted from 1×106 cultured NthyOri 3.1 and FTC133 cells using boiling water and subjected to LC–MS/MS analysis using AS-11 column (Dionex, 2.1 × 250 mm). 13C-fumarate was used as an internal standard. Relative succinate levels were normalized by calculating the ratio of succinate to sum of all measured metabolites as described previously (Denkert et al. 2008).
Oxygen consumption rate (OCR)
FTC133 and NthyOri 3.1 cells were plated at a density of 20,000 cells per well (XF24 cell culture microplate; Seahorse Biosciences, North Billerica, MA, USA). The cells were allowed to grow for 24 h, following which the cells were washed with XF Assay media (with 20 mM glucose, 1 mM sodium pyruvate, no sodium bicarbonate at pH 7.4). The cells were incubated for 1 h at 37°C in a non-CO2 incubator. The assay was normalized to protein and analyzed using the XFe 2.3 software. Optimal seeding density and concentrations of oligomycin, rotenone and FCCP (purchased from Sigma) were determined for each cell line. For the mitochondrial stress test, OCR was measured with sequential addition of oligomycin, rotenone and FCCP. Spare reserve capacity was calculated as FCCP induced-maximum OCR relative to baseline OCR, whereas non-mitochondrial respiration was calculated from the mitochondrial stress test based on residual respiration in response to rotenone.
Cell proliferation/growth
Five thousand cells were plated in 48-well plates for a growth curve by crystal violet assay. At each time point, triplicate wells were stained with 0.05% crystal violet, 0.1% formalin for 20 min and extracted with 10% acetic acid. Absorbance was measured at 590 nm.
Cell migration assay
Cell migration was assessed by scratch wound healing assay. For the measurement of cell migration, control and SDHD-knockdown cells were cultured in individual wells of a 6-well plate. After reaching a confluent state, cell layers were scratched with a 200-µL plastic micropipette tip. The medium was aspirated away and replaced by 1–2 mL of fresh serum-free medium. Cells were allowed to migrate in serum-free medium for 24 h. Images were obtained at 0 and 24 h by phase contrast microscopy. For evaluation of scratch closure, the horizontal distance of migrating cells from the initial wound was measured at 2 points along each scratch using ImageJ software.
Immunohistochemistry
Dissected mouse thyroid tissues were fixed in 10% neutral-buffered formalin solution. Tissues were processed, embedded in paraffin, cut in 5 μm sections on positively charged slides, deparaffinized, rehydrated and stained with H&E. For immunohistochemistry, all sections were stained using a Bond Rx autostainer (Leica Biosystems) for Ki67 (Abcam) antibody. For quantification of the DAB staining, the images were reviewed and analyzed using immunoRatio application.
Electron microscopy
Mouse thyroid tissue was fixed in 2.5% Gluteraldehyde and 0.1 M Phosphate Buffer at pH 7.4 for TEM microscopy. The tissue sections were imaged on FEI Tecnai G2 Spirit TEM microscope.
RNA and real-time PCR
Total RNA from cells and tissue was isolated using Trizol using the Qiagen RNeasy kit. RNA quality was assessed by Nanodrop ND-1000 (Thermo Scientific). cDNA was prepared using an iScript cDNA Synthesis Kit (BioRad Laboratories) and subject to qRT-PCR using the iQ SYBR Green Supermix Kit (BioRad) as per manufacturer's instructions. RT-PCR reactions were performed in triplicate. Primer sequences are available on request.
Genomic DNA isolation
Genomic DNA was isolated from control and SDHD-knockdown cells, and mouse thyroids using DNeasy Blood & Tissue Kit (Qiagen) according to the manufacturer’s protocol. The concentration and purity were determined by measuring the absorbance at 230, 260 and 280 nm using a Nanodrop ND-1000 (Thermo Scientific).
Methylation analysis
Methylated DNA was quantified using 100 ng DNA using an ELISA assay for methylated DNA according to the manufacturer’s directions (Abcam). Absolute and relative methylcytosine content was then calculated using the supplied formula.
Statistics
All data were analyzed via Student’s t-test, Mann–Whitney test or ANOVA using GraphPad Prism software, P-values less than 0.05 were considered significant.
Results
Sdhd knockout causes thyroid hyperplasia in mice
To gain insight into the role of Sdhd in thyroid tumorigenesis, we crossed mice carrying a conditional null allele of Sdhd (SdhdloxP) to mice expressing Cre recombinase under the control of the Tpo promoter in order to generate thyroid-specific Sdhd-KO mice (denoted as Sdhd-TpoKO). Sdhd-TpoKO mice were viable and fertile with normal lifespan. They also demonstrated normal thyroid function, with normal levels of serum TSH (WT: 71.8 ± 43.36 ng/mL vs Sdhd-TpoKO: 57 ± 83.85 ng/mL, P value = 0.7) and no change in the expression of the sodium–iodine symporter mRNA compared to controls (data not shown). Sdhd deletion resulted in a modest increase in thyroid volumes of Sdhd-TpoKO mice at 6 months compared to controls (Fig. 1A), which was sustained up to 12 months (P value for the longitudinal analysis = 0.0075). Histopathologic analysis in a subset of mice revealed incidence of follicular adenoma as 18% (2/11) in Sdhd-TpoKO compared to 0% (0/7) in controls. However, the Sdhd-KO thyroids were more cellular, exhibiting both a decrease in follicular area and a 2-fold enhanced rate of proliferation as measured by Ki-67 staining (Fig. 1B and C). Inflammatory component as well as apoptosis rate was low and unchanged in the Sdhd-TpoKO compared to littermate controls (data not shown). An ultrastructure analysis of the thyroids revealed severe degeneration of the mitochondria in Sdhd-TpoKO mice, similar to previous reports (Szarek et al. 2015), suggesting altered bioenergetics (Fig. 1D).
Sdhd deletion leads to mitochondrial defects and enhanced cellularity in vivo. (A) Thyroid volumes of WT (n = 12) and Sdhd-TpoKO (n = 16) mice analyzed by 3D ultrasonography at 6 months age. (B) Representative images of follicles in WT, Sdhd-TpoKO, Pten-TpoKO and SP-TpoKO mice (200× magnification) at 1 year of age. Quantification of the average follicular size of WT thyroids compared to Sdhd-TpoKO, and Pten-TpoKO thyroids compared to SP-TpoKO is on right. (C) Ki-67 staining by immunohistochemistry in formalin fixed thyroid sections of WT, Sdhd-TpoKO, Pten-TpoKO and SP-TpoKO mice (100× magnification). Quantification of Ki-67 positive nuclei is on right. (D) Mitochondrial ultrastructure (17,000× magnification) of WT and Sdhd-TpoKO mice (n = 3 each) at 6 months of age. Representative TEM images are shown and mitochondria are indicated by the asterisk. Error bars represents standard deviation (s.d.). Statistical analyses were performed by two-tailed Student’s t-test (*P value ≤ 0.05, **P value ≤ 0.01).
Citation: Endocrine-Related Cancer 24, 11; 10.1530/ERC-17-0229
Similarly, to test whether Sdhd deletion enhances tumorigenesis of the thyroid in cooperation with the loss of Pten, we generated Sdhd Pten-TpoKO double knockout mice (denoted as SP-TpoKO) and compared them to mice lacking Pten in their thyroids. We have previously shown that tissue-specific ablation of Pten alone generates mice with enlarged thyroids and follicular adenomas without cancer (Pringle et al. 2014). At 1 year of age, mice with both Sdhd and Pten deletion in the thyroid showed no significant difference from Pten-only deletion in terms of thyroid size or the absence of follicular carcinoma (data not shown). At the cellular level, SP-TpoKO thyroids also showed markedly enhanced proliferation and a reduced follicle size, leading to overall enhanced cellularity in the thyroid, although gross size was not altered (Fig. 1B and C). However when aged up to 18 months, SP-TpoKO mice demonstrated 81% penetrance of follicular carcinomas, whereas Pten-TpoKO did not undergo malignant transformation. Lung metastases were also observed in SP-TpoKO mice. Pten-TpoKO mice were euthyroid (mean TSH 39.2 ± 67.98 ng/mL) and SP-TpoKO mice also exhibited no change in thyroid function (mean TSH 42.2 ± 23.31 ng/mL) (Yeager et al. 2007). Together, these data suggest that Sdhd deletion causes thyroid hyperproliferation as an early sign of tumor initiation, and when combined with Pten KO, has the potential to progress to advanced disease over time.
SDHD deficiency leads to metabolic defects
Because mouse thyroid cells are difficult to manipulate in primary culture, we selected human thyroid cell lines and used lentiviral shRNAs to generate stable SDHD knockdown using Pten-null FTC133 follicular thyroid cancer cell line and the non-malignant NthyOri 3.1 cells as model systems. The efficiency of SDHD knockdown was tested by quantitative RT-PCR and verified by Western blot (Fig. 2A, B and C). Interestingly, SDHD-deficient cells showed normal expression levels of genes encoding the other SDH subunits (data not shown), indicating a lack of co-regulation of the genes encoding this multi-subunit protein. To assess the metabolic effects of SDHD knockdown, we performed quantitative mass spectrometry of small organic acids from the cells. This analysis demonstrated that SDHD downregulation led to an accumulation of succinate in both cell lines following transduction (Fig. 2D and E). Changes in other metabolites of the TCA cycle were inconsistent in the two cell lines (Supplementary Fig. 1, see section on supplementary data given at the end of this article).
SDHD depletion leads to succinate accumulation in vitro. (A) Quantification of SDHD knockdown in FTC133 by RT-PCR. Gene expression is normalized to B2M. (B) Quantification of SDHD knockdown in NthyOri 3.1 by RT-PCR. Gene expression is normalized to β2-microglobulin (B2M). (C) Western blot showing SDHD knockdown in lentivirus-transduced thyroid cell lines in FTC133 and NthyOri 3.1. (D) Quantification of relative succinate levels in FTC133 cells analyzed by mass spectrometry. Data represent average of four biological replicates. (E) Quantification of relative succinate levels in NthyOri 3.1 cells analyzed by mass spectrometry. Data represent average of four biological replicates. Error bars represent standard deviation (s.d.). Statistical analyses were performed by two-tailed Student’s t-test (*P value ≤ 0.05, **P value ≤ 0.01).
Citation: Endocrine-Related Cancer 24, 11; 10.1530/ERC-17-0229
As SDHD is involved in metabolism through its role in TCA cycle and associated ETC, we investigated whether SDHD dysfunction results in defects in cellular respiration and metabolism by Seahorse assay. Interestingly, SDHD knockdown did not completely abrogate the basal OCR in either cell lines. However, SDHD knockdown cell lines responded differently to FCCP treatment (Fig. 3A and B). SDHD depletion thus led to decreased spare capacity and rendered cells more sensitive to energy stress (Fig. 3E and F). Moreover, SDHD-deficient cells showed less sensitivity in response to oligomycin treatment, indicating a greater reliance on non-mitochondrial respiration. SDHD knockdown cells had increased non-mitochondrial respiration than control cells in both cell lines (Fig. 3C and D). Overall, knockdown of SDHD decreased mitochondrial reserve respiratory capacity.
SDHD-KD leads to lower respiration and decreased spare capacity. (A) Seahorse respiratory profile analysis of SDHD-deficient FTC133 cells. (B) Seahorse respiratory profile analysis of SDHD-deficient NthyOri 3.1 cells. (C, D, E and F) SDHD knockdown cells show a corresponding increase in non-mitochondrial respiration and decrease in spare capacity. Data represent one of three representative experiments. Error bars represent standard deviation (s.d.). Statistical analyses were performed by two-tailed Student’s t-test (*P value ≤ 0.05, **P value ≤ 0.01).
Citation: Endocrine-Related Cancer 24, 11; 10.1530/ERC-17-0229
SDHD depletion leads to increased migration in thyroid cancer cell lines
To assess the effect of SDHD knockdown on cellular markers of cancer behavior, we first measured cell proliferation. Over the course of 4 days of growth, there was no difference in cell numbers between WT and SDHD-KD cells for either cell line (Fig. 4A and B). Next, we measured migratory ability using a 2-D scratch assay. In both cell lines, SDHD-KD cells demonstrated an increased migratory capacity compared to control cells (Fig. 4C and D). This migratory phenotype was also confirmed by Boyden chamber assay (Fig. 4E and F).
SDHD depletion does not affect proliferation but increases migratory capacity of thyroid-derived cell lines. (A and B) Proliferation of WT or Sdhd-knockdown FTC133 and NthyOri 3.1 cells analyzed by crystal violet assay. (C and D) Effect of Sdhd knockdown on FTC133 and NthyOri 3.1 cell migration at 24 h as measured by wound healing assays. (E and F) Effect of Sdhd knockdown on FTC133 and NthyOri 3.1 cell migration at 24 h as measured by a Boyden chamber assay. Each graph represents data from three independent experiments. Error bars represent standard deviation (s.d.). Statistical analyses were performed by two-tailed Student’s t-test (*P value ≤ 0.05, **P value ≤ 0.01, ns = non-significant).
Citation: Endocrine-Related Cancer 24, 11; 10.1530/ERC-17-0229
SDHD depletion promotes stemness in thyrocytes
Collectively, our data suggest that loss of SDHD leads to increased cellularity in vivo, and a migratory phenotype in vitro. SDHD dysfunction also leads to altered metabolic properties, mainly reduced oxidative phosphorylation (OXPHOS) in vitro. These characteristics are associated with a so-called stemness phenotype (Pardal et al. 2005, Hermann et al. 2007, Burgess et al. 2014, Li et al. 2017). We further tested if SDHD dysfunction had any impact on cellular differentiation. We utilized our SDHD knockdown human thyroid cancer cell lines to test the expression of stem cell transcription factors. This analysis demonstrated that transcription factors Nanog and Oct-4 were upregulated in SDHD knockdown NthyOri 3.1 cells relative to the control cells (Fig. 5A). We next analyzed the cells for the expression of intracellular ALDH activity, which has been proposed to play a key role in stem/progenitor cell expansion and differentiation as well as tumor initiation and progression (Rodriguez-Torres & Allan 2016). In agreement with RT-PCR data, we observed a marked increase in the fraction of ALDH-positive cells in SDHD-depleted NthyOri 3.1 cells (Fig. 5B). Unlike NthyOri 3.1 cells, FTC133 cells showed no ALDH activity nor increased ‘stem’ gene expression, as these cells are known to lack ALDH activity (Nagayama et al. 2016).
SDHD-KD leads to increased stem-like response in vitro. (A) Quantitative gene expression analysis in NthyOri 3.1 cells. Gene expression is normalized to B2M. Graph represents one of three representative experiments. (B) ALDH activity in SDHD-depleted NthyOri 3.1 cells by ALDEFLUOR assay. The percentage of ALDHhigh cells from single experiment is shown in each panel. On the right, mean of four independent experiments measuring % ALDH-positive cells gated according to DEAB-treated negative controls is quantified. Error bars represent standard deviation (s.d.). Statistical analyses were performed by two-tailed Student’s t-test (*P value ≤ 0.05, **P value ≤ 0.01).
Citation: Endocrine-Related Cancer 24, 11; 10.1530/ERC-17-0229
With the identification of stem-like features in the cell lines, we wanted to see if the same observations were applicable to murine tumors in vivo, as acquisition of this stem-like phenotype would be a significant indicator of tumor initiation and neoplasia. To test this hypothesis, we isolated P and SP thyrocytes for measurement of their ALDH activity. As shown in Fig. 6, the population of ALDH-positive cells in SP thyroid tumors was significantly elevated compared to Pten-only mice. These data confirm that the ability of Sdhd ablation to promote tumorigenesis is significantly reliant on the promotion of this stem-like phenotype.
SDHD-KD leads to increased stemness response in vivo. (A) ALDH activity in primary mouse thyrocytes by ALDEFLUOR assay. The percentage of ALDHhigh cells from single experiment is shown in each panel. (B) Mean of three independent experiment measuring % ALDH-positive cells gated according to DEAB-treated negative control. Error bars represent standard deviation (s.d.). Statistical analyses were performed by two-tailed Student’s t-test (*P value ≤ 0.05).
Citation: Endocrine-Related Cancer 24, 11; 10.1530/ERC-17-0229
SDHD deficiency increases global DNA methylation in thyroid cells
As described earlier, three mechanisms have been proposed to account for the tumorigenic effect of mutation of genes in the TCA cycle: generation of free radicals, hypoxia/pseudohypoxia and epigenetic alterations caused by metabolite excess. To assess free radical generation, we measured ROS in control and SDHD-deficient cells; while ROS levels could be manipulated by altering the oxidizing environment of the cells, no differences were detected between control and KD cells at baseline or under stimulation (Supplementary Fig. 2). To produce a pseudohypoxic state, it has been suggested that succinate accumulation can inhibit prolyl hydroxylase leading to a hypoxic response (Briere et al. 2005). However, we did not observe induction of HIF-1α or -2α (data not shown), suggesting that this pathway is not operative in these cells. We also examined the ERK, PI3K and mTOR pathways to evaluate activation in SDHD-deficient cells. While p-AKT and p-ERK levels remained unchanged both in vivo and in vitro (data not shown), we observed mildly increased levels of p-mTOR in Sdhd-TpoKO and SP-TpoKO mouse thyroids. However, the ratio of p-mTOR to total mTOR was unaltered in vitro (Supplementary Fig. 3).
It has been shown that TCA cycle intermediates fumarate and succinate can inhibit the histone demethylases of the Jumonji C class in a dose-dependent manner (Xiao et al. 2012). Also, SDHC mutations give rise to hypermethylator phenotype in PGL (Letouze et al. 2013). Thus, we tested whether SDHD depletion directly affected DNA methylation in thyroid cell lines and mouse tumors. In NthyOri 3.1 cells, there was an approximately 3-fold increase in 5-methylcytosine (5-mC) levels after depletion of SDHD (Fig. 7A). In contrast, DNA methylation levels in FTC133 cells seemed to be unaffected by SDHD loss. It is unclear if the lack of change is similar to (or caused by) the same defect that prevents ALDH expression, or if this is a limitation of the sensitivity of the assay (Fig. 7B). In mouse tumors, there was a clear trend toward an increase in 5-mC levels between SP tumors and Pten-only tumors although these results did not reach significance, perhaps due to high variability observed among mice (Fig. 7C).
DNA methylation analysis of SDHD-depleted cells. (A) 5-mC levels in NthyOri 3.1 cells DNA analyzed by ELISA assay. Graph represents mean ± s.d. of three biological replicates. (B) 5-mC levels in FTC133 cells DNA analyzed by ELISA assay. Graphs represent mean ± s.d. of three biological replicates. (C) 5-mC levels in DNA isolated from SP-TpoKO and Pten-TpoKO tumors analyzed by ELISA assay. Graphs represent mean ± s.d. of five mice. Statistical analyses were performed by two-tailed Student’s t-test (*P value ≤ 0.05, ns = non-significant).
Citation: Endocrine-Related Cancer 24, 11; 10.1530/ERC-17-0229
α-ketoglutarate treatment reverses stem cell-like phenotype caused by mutant SDHD
It is increasingly appreciated that DNA methylation has an important role in cancer development and it helps to maintain transcriptional silencing of genes (Baylin 2005). Epigenetic modifications are reversible and can be therapeutically targeted. The phenotype caused by SDHD loss can be explained by possibly altered αKG/succinate ratio, as we have reported increased succinate levels in our model. If an alteration of the relative amounts of TCA cycle intermediates is responsible for this effect, then reestablishing the balance should revert this effect. To test this possibility, we treated cells with cell-permeable α-KG and measured the phenotypic characteristics as described earlier. Indeed, treatment with low dose of α-KG for three days fully reversed the migratory phenotype as well as stem-like response evidenced by ALDH assay and reduced the ALDH levels back to that of control cells (Fig. 8A and C). To examine whether unavailability of α-KG is also responsible for increased levels of DNA methylation observed with SDHD deficiency, we measured 5-mC levels in NthyOri 3.1 cells treated with α-KG. Indeed, treatment with a high dose of α-KG could rescue the hypermethylation phenotype in NthyOri cells (Fig. 8B).
α-KG reverses stem-like phenotype in vitro. (A) SDHD knockdown NthyOri 3.1 cells with aKG (0.1 mM) treatment for 3 days were subjected to FACS analysis for measurement of ALDH activity. The percentage of ALDHhigh cells from single experiment is shown in each panel. Quantification of three independent experiments measuring % ALDH-positive cells gated according to DEAB-treated negative control is shown on the right (statistical comparison is for the untreated Sh-Ctrl). (B) DNA methylation levels of NthyOri 3.1 cells after treatment with 1 mM α-KG for 3 days. Graphs represent mean ± s.d. of five biological replicates. (C) Migration in thyroid cell lines treated with aKG (0.1 mM) at 12 h performed by wound healing assays. Graph represents data from three independent experiments. Error bars represent standard deviation (s.d.). Statistical analyses were performed by two-tailed Student’s t-test for ALDH assay and migration, and by Mann–Whitney test in the case of DNA methylation (*P value ≤ 0.05, **P value ≤ 0.01).
Citation: Endocrine-Related Cancer 24, 11; 10.1530/ERC-17-0229
Discussion
Although mutations in SDH genes are known to cause human cancers, analysis of the mechanism by which this occurs has been hampered by a lack of robust model systems. Previous efforts to study tumorigenesis in vivo did not identify an effect of mutant IDH, FH or SDH genes, including in tissues clearly affected in human patients (Lu et al. 2013, Millan-Ucles et al. 2014, Szarek et al. 2015). In this report, we address the role of SDHD in thyroid cancer by ablating Sdhd in the thyroid gland. Because it has been suggested that there may be an interaction between loss of the PTEN tumor suppressor and mutations in SDH genes, we also generated mice lacking both these genes and compared them to Pten KO thyroids, which only develop thyroid adenomas. Studying the role of tumor suppressors and oncogenes in the thyroid provides an advantageous model system to study pre-neoplastic and neoplastic changes because of the unique architecture of the thyroid follicle, as well as the opportunity to study the differentiated function of thyroid hormone secretion.
In contrast to the prior studies looking at the role of SDHD in neural crest-derived tissue (including the adrenal medulla), we observed that Sdhd null thyroids show enlarged glands, a potential indicator of early neoplasia. Examination of the tissue revealed hypercellularity caused by enhanced proliferation, further supporting the impression that Sdhd ablation promotes tumor initiation in the thyroid gland. In fact, when combined with Pten KO, these mice developed FTC with metastasis at advanced age. Although Sdhd ablation from the thyroid causes enhanced proliferation in vivo, we did not observe this same effect in vitro. We attribute the differences either to the fact that in vitro studies were carried out in permanent cell lines, which may already have altered proliferation properties, or to the lack of a normal 3D tissue environment which exists in vivo.
One of the most striking observations from this study is that KO of Sdhd from the thyroid in vivo or knockdown of the gene in vitro promotes a shift to a stem-like phenotype, including the expression of both stem cell transcription factors and cellular production of the thyroid stem cell marker ALDH. Although the tumor cells retained sufficiently differentiated function to maintain normal thyroid function, the acquisition of stem-like features is likely to play an important role in the ability of the hyperplastic glands to eventually progress to thyroid cancer.
Three mechanisms have been proposed to account for the mechanism by which mutations in SDH (or other TCA cycle genes) may account for the acquisition of a malignant phenotype. In the thyroid gland, we could find no evidence for an excess of ROS nor for significant induction of a pseudohypoxia signature. These results differ from a number of studies on TCA cycle tumor suppressors that have resulted in inconsistent reports on HIF and ROS involvement, casting doubts on pseudohypoxia and free radical mechanisms (Briere et al. 2005, Aspuria et al. 2014, Millan-Ucles et al. 2014, Edalat et al. 2015). Based on our observations, we focused on the role of succinate as an oncometabolite. Levels of succinate were elevated in SDHD-KD cells, although there were minimal effects on other TCA cycle intermediates. Succinate has been proposed to affect DNA demethylases and, indeed, knockdown cells exhibited increased DNA methylation, an observation which was seen in vivo but which did not reach statistical significance due to high variability. The in vivo studies may reflect not only variability in the thyrocytes themselves, but also contributions from stromal cells, which would be expected to have normal SDH function and thus reduce the magnitude of any alterations restricted to thyroid epithelial cells.
It has been proposed that the effects of oncometabolites are mediated by the imbalance, which alters the ratio of various allosteric enzyme modulators. In the case of succinate, it has been suggested that the effects are mediated by an imbalance in the ratio of succinate to its precursor, alpha-ketoglutarate (α-KG). We tested this experimentally by treating cells with α-KG and demonstrated a nearly complete reversal of phenotypes with this treatment. The fact that SDH-associated changes can be reversed by excess α-KG suggests that altered α-KG/succinate ratio, rather than absolute levels, likely contributes to tumorigenesis by enzymatic inhibition of α-KG-dependent dioxygenases. The differential response of SDH dysfunction in different cell systems may be attributed to the strength of epigenetic effects which often varies between cell types. This observation may have clinical implications, although one would need to be cautious about manipulation of these metabolites, lest other intracellular processes were disturbed.
As expected, Sdhd-KO thyroids exhibited aberrant mitochondria, and studies in vitro demonstrated altered respiratory function, consistent with the metabolic remodeling. In the case of SDHD mutations, our data are consistent with the notion that the metabolic abnormalities can drive tumorigenesis, rather than occur as a secondary effect. The thyroid gland is a unique environment in that it is a tissue that is highly reliant on intracellular oxidation (through the action of Tpo) for its biological function of generating thyroid hormone. This fact may render thyrocytes either susceptible or more resistant to changes that affect ROS. In our hands, we did not observe a noticeable effect on ROS.
In summary, Sdhd knockout from the thyroid is sufficient to cause excess thyroid cell growth in mice, which may play a role in tumor initiation. In addition, the present results unveil a role for an aberrant TCA cycle in the generation of a stem-like phenotype in vivo and in vitro. Together, our data tie metabolic dysfunction with tumorigenic response and warrant further studies for in-depth analysis of regulation of gene expression due to epigenetic changes in each model. These data also suggest that therapeutic reversal of DNA methylation may arise as an attractive approach to add to existing treatments of SDHx-mutated tumors. The identification of a hypermethylator phenotype, albeit observed in single non-tumor derived NthyOri cell line, helps in explaining both the tumor-suppressive role of SDH and the context-specific phenotypic characteristics. It warrants further studies of the impact of these mutations on specific genomic regions which will help us to identify signaling pathways that play a role in the SDH-related oncogenesis.
Supplementary data
This is linked to the online version of the paper at http://dx.doi.org/10.1530/ERC-17-0229.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
This work was supported by NIH grants P01CA124570 (to LSK) and P30CA0168058 (OSU Comprehensive Cancer Center). AA, DH and AM were supported in part by individual Pelotonia Fellowship Program grants.
Acknowledgements
The authors would like to thank Dr Jean-Christophe Cocuron and the OSU Targeted Metabolomics Laboratory for the mass spectrometry analysis of cellular metabolites. The authors would also like to acknowledge Dr Xuguang Zhu and Shueu-yann Cheng (Laboratory of Molecular Biology, National Cancer Institute, Bethesda, MD) for performing TSH measurements on the mouse serum. Dr Matthew Ringel and Dr Charis Eng provided insightful comment on this work during its development.
References
Archetti M 2015 Heterogeneity and proliferation of invasive cancer subclones in game theory models of the Warburg effect. Cell Proliferation 48 259–269. (doi:10.1111/cpr.12169)
Aspuria PJ, Lunt SY, Varemo L, Vergnes L, Gozo M, Beach JA, Salumbides B, Reue K, Wiedemeyer WR & Nielsen J et al. 2014 Succinate dehydrogenase inhibition leads to epithelial-mesenchymal transition and reprogrammed carbon metabolism. Cancer and Metabolism 2 21. (doi:10.1186/2049-3002-2-21)
Bardella C, Pollard PJ & Tomlinson I 2011 SDH mutations in cancer. Biochimica et Biophysica Acta 1807 1432–1443. (doi:10.1016/j.bbabio.2011.07.003)
Baylin SB 2005 DNA methylation and gene silencing in cancer. Nature Clinical Practice Oncology 2 (Supplement 1) S4–S11. (doi:10.1038/ncponbib354)
Briere JJ, Favier J, Benit P, El Ghouzzi V, Lorenzato A, Rabier D, Di Renzo MF, Gimenez-Roqueplo AP & Rustin P 2005 Mitochondrial succinate is instrumental for HIF1alpha nuclear translocation in SDHA-mutant fibroblasts under normoxic conditions. Human Molecular Genetics 14 3263–3269. (doi:10.1093/hmg/ddi359)
Burgess RJ, Agathocleous M & Morrison SJ 2014 Metabolic regulation of stem cell function. Journal of Internal Medicine 276 12–24. (doi:10.1111/joim.12247)
Cantor JR & Sabatini DM 2012 Cancer cell metabolism: one hallmark, many faces. Cancer Discovery 2 881–898. (doi:10.1158/2159-8290.CD-12-0345)
Carey BW, Finley LW, Cross JR, Allis CD & Thompson CB 2015 Intracellular alpha-ketoglutarate maintains the pluripotency of embryonic stem cells. Nature 518 413–416. (doi:10.1038/nature13981)
Davies L, Morris LG, Haymart M, Chen AY, Goldenberg D, Morris J, Ogilvie JB, Terris DJ, Netterville J & Wong RJ et al. 2015 American Association of Clinical Endocrinologists and American College of Endocrinology disease state clinical review: the increasing incidence of thyroid cancer. Endocrine Practice 21 686–696. (doi:10.4158/EP14466.DSCR)
Denkert C, Budczies J, Weichert W, Wohlgemuth G, Scholz M, Kind T, Niesporek S, Noske A, Buckendahl A & Dietel M et al. 2008 Metabolite profiling of human colon carcinoma – deregulation of TCA cycle and amino acid turnover. Molecular Cancer 7 72. (doi:10.1186/1476-4598-7-72)
Edalat A, Schulte-Mecklenbeck P, Bauer C, Undank S, Krippeit-Drews P, Drews G & Dufer M 2015 Mitochondrial succinate dehydrogenase is involved in stimulus-secretion coupling and endogenous ROS formation in murine beta cells. Diabetologia 58 1532–1541. (doi:10.1007/s00125-015-3577-9)
Hermann PC, Huber SL, Herrler T, Aicher A, Ellwart JW, Guba M, Bruns CJ & Heeschen C 2007 Distinct populations of cancer stem cells determine tumor growth and metastatic activity in human pancreatic cancer. Cell Stem Cell 1 313–323. (doi:10.1016/j.stem.2007.06.002)
Hobert JA & Eng C 2009 PTEN hamartoma tumor syndrome: an overview. Genetics in Medicine 11 687–694. (doi:10.1097/GIM.0b013e3181ac9aea)
Jardim-Messeder D, Caverzan A, Rauber R, de Souza Ferreira E, Margis-Pinheiro M & Galina A 2015 Succinate dehydrogenase (mitochondrial complex II) is a source of reactive oxygen species in plants and regulates development and stress responses. New Phytologist 208 776–789. (doi:10.1111/nph.13515)
Kitahara CM & Sosa JA 2016 The changing incidence of thyroid cancer. Nature Reviews Endocrinology 12 646–653. (doi:10.1038/nrendo.2016.110)
Kusakabe T, Kawaguchi A, Kawaguchi R, Feigenbaum L & Kimura S 2004 Thyrocyte-specific expression of Cre recombinase in transgenic mice. Genesis 39 212–216. (doi:10.1002/gene.20043)
Leenhardt L, Grosclaude P & Cherie-Challine L 2004 Increased incidence of thyroid carcinoma in france: a true epidemic or thyroid nodule management effects? Report from the French Thyroid Cancer Committee. Thyroid 14 1056–1060. (doi:10.1089/thy.2004.14.1056)
Letouze E, Martinelli C, Loriot C, Burnichon N, Abermil N, Ottolenghi C, Janin M, Menara M, Nguyen AT & Benit P et al. 2013 SDH mutations establish a hypermethylator phenotype in paraganglioma. Cancer Cell 23 739–752. (doi:10.1016/j.ccr.2013.04.018)
Li Q, Ye L, Zhang X, Wang M, Lin C, Huang S, Guo W, Lai Y, Du H & Li J et al. 2017 FZD8, a target of p53, promotes bone metastasis in prostate cancer by activating canonical Wnt/beta-catenin signaling. Cancer Letters 402 166–176. (doi:10.1016/j.canlet.2017.05.029)
Lu C, Venneti S, Akalin A, Fang F, Ward PS, Dematteo RG, Intlekofer AM, Chen C, Ye J & Hameed M et al. 2013 Induction of sarcomas by mutant IDH2. Genes and Development 27 1986–1998. (doi:10.1101/gad.226753.113)
MacKenzie ED, Selak MA, Tennant DA, Payne LJ, Crosby S, Frederiksen CM, Watson DG & Gottlieb E 2007 Cell-permeating alpha-ketoglutarate derivatives alleviate pseudohypoxia in succinate dehydrogenase-deficient cells. Molecular and Cellular Biology 27 3282–3289. (doi:10.1128/MCB.01927-06)
Marsh DJ, Coulon V, Lunetta KL, Rocca-Serra P, Dahia PL, Zheng Z, Liaw D, Caron S, Duboue B & Lin AY et al. 1998 Mutation spectrum and genotype-phenotype analyses in Cowden disease and Bannayan-Zonana syndrome, two hamartoma syndromes with germline PTEN mutation. Human Molecular Genetics 7 507–515. (doi:10.1093/hmg/7.3.507)
Millan-Ucles A, Diaz-Castro B, Garcia-Flores P, Baez A, Perez-Simon JA, Lopez-Barneo J & Piruat JI 2014 A conditional mouse mutant in the tumor suppressor SdhD gene unveils a link between p21(WAF1/Cip1) induction and mitochondrial dysfunction. PLoS ONE 9 e85528. (doi:10.1371/journal.pone.0085528)
Morrison PJ & Atkinson AB 2009 Genetic aspects of familial thyroid cancer. Oncologist 14 571–577. (doi:10.1634/theoncologist.2009-0046)
Nagayama Y, Shimamura M & Mitsutake N 2016 Cancer stem cells in the thyroid. Frontiers in Endocrinology 7 20. (doi:10.3389/fendo.2016.00020)
Nagy R, Sweet K & Eng C 2004 Highly penetrant hereditary cancer syndromes. Oncogene 23 6445–6470. (doi:10.1038/sj.onc.1207714)
Ni Y, Zbuk KM, Sadler T, Patocs A, Lobo G, Edelman E, Platzer P, Orloff MS, Waite KA & Eng C 2008 Germline mutations and variants in the succinate dehydrogenase genes in Cowden and Cowden-like syndromes. American Journal of Human Genetics 83 261–268. (doi:10.1016/j.ajhg.2008.07.011)
Ni Y, Seballos S, Ganapathi S, Gurin D, Fletcher B, Ngeow J, Nagy R, Kloos RT, Ringel MD & LaFramboise T et al. 2015 Germline and somatic SDHx alterations in apparently sporadic differentiated thyroid cancer. Endocrine-Related Cancer 22 121–130. (doi:10.1530/ERC-14-0537)
Nose V 2011 Familial thyroid cancer: a review. Modern Pathology 24 (Supplement 2) S19–S33. (doi:10.1038/modpathol.2010.147)
Pardal R, Molofsky AV, He S & Morrison SJ 2005 Stem cell self-renewal and cancer cell proliferation are regulated by common networks that balance the activation of proto-oncogenes and tumor suppressors. Cold Spring Harbor Symposia on Quantitative Biology 70 177–185. (doi:10.1101/sqb.2005.70.057)
Pringle DR, Yin Z, Lee AA, Manchanda PK, Yu L, Parlow AF, Jarjoura D, La Perle KM & Kirschner LS 2012 Thyroid-specific ablation of the Carney complex gene, PRKAR1A, results in hyperthyroidism and follicular thyroid cancer. Endocrine-Related Cancer 19 435–446. (doi:10.1530/ERC-11-0306)
Pringle DR, Vasko VV, Yu L, Manchanda PK, Lee AA, Zhang X, Kirschner JM, Parlow AF, Saji M & Jarjoura D et al. 2014 Follicular thyroid cancers demonstrate dual activation of PKA and mTOR as modeled by thyroid-specific deletion of Prkar1a and Pten in mice. Journal of Clinical Endocrinology and Metabolism 99 E804–E812. (doi:10.1210/jc.2013-3101)
Rodriguez-Torres M & Allan AL 2016 Aldehyde dehydrogenase as a marker and functional mediator of metastasis in solid tumors. Clinical and Experimental Metastasis 33 97–113. (doi:10.1007/s10585-015-9755-9)
Salminen A, Kaarniranta K, Hiltunen M & Kauppinen A 2014 Krebs cycle dysfunction shapes epigenetic landscape of chromatin: novel insights into mitochondrial regulation of aging process. Cell Signaling 26 1598–1603. (doi:10.1016/j.cellsig.2014.03.030)
Sciacovelli M & Frezza C 2016 Oncometabolites: unconventional triggers of oncogenic signalling cascades. Free Radical Biology and Medicine 100 175–181. (doi:10.1016/j.freeradbiomed.2016.04.025)
Szarek E, Ball ER, Imperiale A, Tsokos M, Faucz FR, Giubellino A, Moussallieh FM, Namer IJ, Abu-Asab MS & Pacak K et al. 2015 Carney triad, SDH-deficient tumors, and Sdhb+/- mice share abnormal mitochondria. Endocrine-Related Cancer 22 345–352. (doi:10.1530/ERC-15-0069)
Tomsic J, He H, Akagi K, Liyanarachchi S, Pan Q, Bertani B, Nagy R, Symer DE, Blencowe BJ & de la Chapelle A 2015 A germline mutation in SRRM2, a splicing factor gene, is implicated in papillary thyroid carcinoma predisposition. Scientific Reports 5 10566. (doi:10.1038/srep10566)
Williamson SR, Eble JN, Amin MB, Gupta NS, Smith SC, Sholl LM, Montironi R, Hirsch MS & Hornick JL 2015 Succinate dehydrogenase-deficient renal cell carcinoma: detailed characterization of 11 tumors defining a unique subtype of renal cell carcinoma. Modern Pathology 28 80–94. (doi:10.1038/modpathol.2014.86)
Xiao M, Yang H, Xu W, Ma S, Lin H, Zhu H, Liu L, Liu Y, Yang C & Xu Y et al. 2012 Inhibition of alpha-KG-dependent histone and DNA demethylases by fumarate and succinate that are accumulated in mutations of FH and SDH tumor suppressors. Genes and Development 26 1326–1338. (doi:10.1101/gad.191056.112)
Yeager N, Klein-Szanto A, Kimura S & Di Cristofano A 2007 Pten loss in the mouse thyroid causes goiter and follicular adenomas: insights into thyroid function and Cowden disease pathogenesis. Cancer Research 67 959–966. (doi:10.1158/0008-5472.CAN-06-3524)
Yu W, He X, Ni Y, Ngeow J & Eng C 2015 Cowden syndrome-associated germline SDHD variants alter PTEN nuclear translocation through SRC-induced PTEN oxidation. Human Molecular Genetics 24 142–153. (doi:10.1093/hmg/ddu425)
