Antibodies and Reagents
BPTES (HY-12683), CBR-5884 (HY-100012), SHIN1 (HY-112066), DS18561882 (HY-130251), and sorafenib (HY-10201) were purchased from MedChemExpress. Trolox (S3665), lenvatinib (S1164) were from selleckchem. Crystal violet (V5265), H2DCFDA (D399), MitoSOXTM Red (M36008), monobromobimane (mBBr, M1378), (3-(4,5-Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium Bromide) (MTT, M6494), and LipofectamineTM RNAiMAX Transfection Reagent (13778150) were obtained from Thermo Fisher Scientific. Primary antibodies against SHMT2 (sc-390641), PHGDH (sc-100317), and β-actin (sc-47778) were purchased from Santa Cruz Biotechnology. MTHFD2 (98116S) and Cell Signaling Technology. Secondary antibodies of mouse and rabbit were from Cell Signaling Technology and abcam respectively (7076S and ab6721). Small interfering RNA (siCTRL and siATF4) oligonucleotides were purchased from Santa Cruz Biotechnology (sc-37007 and sc-35112, respectively).
Cell culture and treatment
8505C cells obtained from European Collection of Authenticated Cell Cultures (ECACC) were grown in DMEM medium (Corning, 10-013-CV)-1% penicillin-streptomycin (Gibco, 15140122) supplemented with 10% fetal bovine serum (Corning, 35-015-CV). Cells were tested for mycoplasma elimination using a Plasmocin solution (Invivogen, ant-mpt). For in vitro treatment, following concentrations were employed: Glutamine-full medium consists of DMEM medium (Welgene, LM-001-05)-1% penicillin-streptomycin (Gibco, 15140122) supplemented with 10% fetal bovine serum (Corning, 35-015-CV); Glutamine-free medium consists of DMEM medium (Welgene, LM-001-08)-1% penicillin-streptomycin (Gibco, 15140122) supplemented with 10% dialyzed fetal bovine serum (Gibco, 26400044); MTT, 0.2 mg/ml; Crystal violet solution, 0.1%; mBBr, 10 uM; H2DCFDA, 10 uM; MitoSOXTM Red, 5 uM; BPTES, 10 uM; CBR-5884, 60 uM; Trolox, 25 uM; lenvatinib, 50 uM; sorafenib, 10 uM.
Cell proliferation and viability assay
8505C cells (5×103 and 1×104 cells/well) were seeded in 96-well plates for proliferation and viability analysis, respectively. Cells were exposed to glutamine-free medium or reagents for the indicated times, with the vehicle as a control. Cells were incubated for 2 h in a 37°C incubator after addition of 10 µl of MTT solution and formazan was dissolved in 50 µl DMSO (Sigma, 34943). The absorbance was measured at 570 nm using a Multiskan GO spectrophotometer (Thermo Fisher Scientific, 51119300).
Colony formation assay
8505C cells (5×103 cells/well) were seeded in 6-well plates and treated with glutamine-free medium, BPTES, for 10 days. After washing with PBS twice, cells were fixed with cold methanol at room temperature for 5 min and stained with crystal violet for 15 min. After crystal violet was slowly removed, the plates were air-dried overnight. Each sample was then added to 1 ml of methanol and rotated for 20 min. The optical density of each well was measured at 570 nm using a Multiskan GO spectrophotometer (Thermo Fisher Scientific, 51119300).
LDH assay
The 8505C cells (5×103 cells/well) were seeded in a 96-well plate. After 24 h, the cells were treated with glutamine-free medium or BPTES for the indicated times. Cell death was assessed based on the release of LDH into the extracellular medium, which was measured using a Cytotoxicity Detection Kit according to the manufacturer’s protocol (Thermo Fisher Scientific, C20301).
Real-time RT-PCR
RNA was isolated using TRIzol reagent (Invitrogen, 15596018). cDNA was synthesized using 1 ug total RNA and ImProm-II™ Reverse Transcriptase (Promega, A3803). Real-time RT-PCR was performed using TOP Real ™ qPCR 2X Pre-MIX (Enzynomics, RT501S) and specific primers in a CFX Connect Real-Time PCR instrument (Bio-Rad, 1855201). Gene expression was normalized to the 36B4 mRNA expression levels.
The primer sequences for real-time RT-PCR are listed in Supplementary Table. 1.
Immunoblot assays
Protein lysates were lysed in mammalian lysis buffer (25 mM Tris-HCl, pH 7.8, 150 mM NaCl, 0.1% NP-40, 1 mM EDTA, 10% glycerol) supplemented with Xpert Protease Inhibitor Cocktail Solution (GenDEPOT, P3100-001). Protein concentrations were quantified using the Bio-Rad Protein Assay Kit (#5000006). Samples were separated by SDS-PAGE (8-10%) and transferred onto nitrocellulose membrane (Amersham, 10600001). Membranes blocked with 5% skim milk-Tris-0.1% Tween 20 for 30 min and incubated diluted antibodies at 4°C overnight. After incubation with HRP-conjugated secondary antibodies for 1 h at room temperature, immunoblot signals were detected using Clarity Western ECL Substrate (BR1705061 and 1705062). For gene silencing with siCTRL or siATF4, the cells were transfected for 48 h with LipofectamineTM RNAiMAX Transfection Reagent according to the manufacturer’s protocol.
Study Design and Ethical Considerations
This study was a retroactive, single center examination of patients diagnosed with ATC (September 2021–January 2022). All courses entailing patients were achieved in proportion to the institutional ethical standards, whole applicable national/local regulations, and guidelines of the 1964 Helsinki Declaration and its later amendments. The study procedures were authorized by the Institutional Review Board (IRB) of Gangnam Severance Hospital, Yonsei University College of Medicine (IRB protocol: 3-2021-0043).
IHC staining
After the samples were deparaffinized and rehydrated, sections for PHGDH staining were placed in FLEX Target solution (DAKO, K8004 (pH9.0)) for antigen retrieval by boiling in a PT link for 20 min at 95°C. For inactivation of endogenous peroxidase, sections were treated with 3% H2O2 for 10 min and washed with TBS for 5 min twice. Next, the slides were incubated with PHGDH antibody (1:100) for 1 h at room temperature. After washing three times with TBS for 5 min each, the slides were incubated with the secondary antibody (DAKO, K4003) for 20 min at room temperature. Diaminobenzidine (DAB) (DAKO, K3468) was used for color development for 5 min. Finally, the slides were counterstained with hematoxylin for 10 min, dehydrated, and mounted. Quantitative data for comparing PHGDH expression in patients were analyzed using Qupath software. (version 0.4.2).
Flow cytometry
The 8505C cells (2×105 cells/well) were seeded in a 12-well plate. After 24 h, the cells were treated with the indicated drugs; vehicle (DMSO), BPTES, CBR-5884, lenvatinib and sorafenib for the indicated times. For intracellular ROS, the cells were stained with H2DCFDA for 30 min at 37°C. For mitochondrial ROS measurement, the cells were stained with MitoSOXTM Red for 30 min at 37°C. To determine the total GSH content, cells were stained with mBBr for 10 min in a 37°C incubator. Fluorescence intensity was quantified by flow cytometry (FACSVerse, BD Biosciences) within 30 min, and the data were analyzed using FlowJo software (version 10.4.2).
RNA sequencing
Total RNA samples were duplicated, and processing was performed by Macrogen Inc. (Seoul, Korea; www.macrogen.com). First, a library was constructed using the TruSeq Stranded mRNA LT Sample Prep Kit (Illumina, San Diego, CA, USA) according to the TruSeq Stranded mRNA Sample Preparation Guide (part #15031047 Rev. E). Next, sequencing was performed following the NovaSeq 6000 System User Guide (Document #1000000019358 v02). The sequence was qualified by FastQC (version 0.11.7), trimmed by Trimmomatic (0.38), and mapped using the HISAT2 (version 2.1.0) program. We assembled gene and transcript expression levels to read counts or fragments per kilobase of transcript per million mapped reads (FPKM) using StringTie (version 2.1.3b). Trimmed mean of M-value (TMM) normalization was performed to reduce systematic bias using read count by the edgeR package library. Finally, the DEGs were estimated using edgeR.
ATAC sequencing
Total RNA samples were duplicated, and processing was performed by Macrogen Inc. (Seoul, Korea; www.macrogen.com). The sequence was qualified using FastQC (version 0.11.7), trimmed using Trim Galore (version 0.5.0), and aligned using the Bowtie2 (version 2.3.5.1) tool. Peak calling from alignment bam files was performed using MACS2 (version 2.1.1.20160309). Raw BAM files were normalized using the Galaxy tool based on the hg38.blacklist.bed file. The total peak signals from the raw bed files are depicted by ShinyCircos. Transcription start site or Functional elements peak signals were analyzed from computeMatrix in Galaxy tool. These two open sources were acquired from UCSC Genome Browser. The peak signals for each gene were visualized using GBiB tool.
Acquisition of scRNA-seq data
We obtained all the findings of single cell analysis from the GEO database. The scRNA-seq data consisting of thyrocytes from ATC (n=5) and PTC patients (n=7) were acquired from GSE148673 and GSE184362 datasets respectively. ‘Seurat (version 4.3.0)’ and ‘DoubletFinder (version 2.0.3)’ packages in R software (version 4.1.3) were used for scRNA-seq data preprocessing. First, scRNA-seq expression matrices were inserted into R using ‘Read10X’. Quality control of each data was performed to filter out poor-quality cells based on numbers of genes per cell (nFeature), whole number of read counts (nCount) and the read counts percentage of mitochondrial genes per cell (percent_MT). After eliminating doublets using ‘doubletFinder_v3’, the scRNA-seq data were normalized by the ‘LogNormalize’, and the top 2000 highly variable genes were identified using ‘FindVariableFeatures’. Integration of each data was performed using ‘IntegrateData’. Information about quality control and percent of doublet are represented in Supplementary Table. 2 in more detail.
Processing of the scRNA-seq data
Cell cycle score of each single cell was calculated using ‘CellCycleScoring’ and integrated data were regressed out to mitigate cell cycle heterogeneity on cell clustering. PCA was utilized with the top fifteen PCA values for dividing clusters using ‘RunPCA’. Total seven major cell types were annotated to each cluster according to marker gene expression through ‘FindAllMarkers’. Cluster visualization was performed using UMAP algorithm (resolution=0.15). ‘DotPlot’ was used to plot marker genes to identify each cell cluster identity. ‘RunALRA’ were performed to recover missing values in cluster of thyrocytes. ‘FeaturePlot’ was used to compare several genes related to cancer type-specific marker and one carbon metabolism between ATC and PTC.
Trajectory inference and pseudotime analysis based on TDS score
To distinguish ATC and PTC in Thyrocyte Cluster, we first calculated TDS score using ‘AddModuleScore’ and further investigated expression of several genes. Next, we assigned starting point of the pseudotime analysis on the basis of TDS score using Monocle3 (version 1.3.1) package and represented scatter plot of several genes along pseudotime from PTC to ATC.
Statistical analysis
Data are expressed as the mean ± SD of three independent experiments. A two-tailed Student's t-test was used to compare the values between the two groups. One-way ANOVA with Tukey's test was used to compare multiple groups. Statistical significance was set at p < 0.05. Data were analyzed using the GraphPad Prism 9 software.