Nicotinamide Nucleotide Transhydrogenase as a Novel Treatment Target in Adrenocortical Carcinoma

Abstract Adrenocortical carcinoma (ACC) is an aggressive malignancy with poor response to chemotherapy. In this study, we evaluated a potential new treatment target for ACC, focusing on the mitochondrial reduced form of NAD phosphate (NADPH) generator nicotinamide nucleotide transhydrogenase (NNT). NNT has a central role within mitochondrial antioxidant pathways, protecting cells from oxidative stress. Inactivating human NNT mutations result in congenital adrenal insufficiency. We hypothesized that NNT silencing in ACC cells will induce toxic levels of oxidative stress. To explore this, we transiently knocked down NNT in NCI-H295R ACC cells. As predicted, this manipulation increased intracellular levels of oxidative stress; this resulted in a pronounced suppression of cell proliferation and higher apoptotic rates, as well as sensitization of cells to chemically induced oxidative stress. Steroidogenesis was paradoxically stimulated by NNT loss, as demonstrated by mass spectrometry–based steroid profiling. Next, we generated a stable NNT knockdown model in the same cell line to investigate the longer lasting effects of NNT silencing. After long-term culture, cells adapted metabolically to chronic NNT knockdown, restoring their redox balance and resilience to oxidative stress, although their proliferation remained suppressed. This was associated with higher rates of oxygen consumption. The molecular pathways underpinning these responses were explored in detail by RNA sequencing and nontargeted metabolome analysis, revealing major alterations in nucleotide synthesis, protein folding, and polyamine metabolism. This study provides preclinical evidence of the therapeutic merit of antioxidant targeting in ACC as well as illuminating the long-term adaptive response of cells to oxidative stress.

system following the Library prep TruSeq Stranded mRNA Library Prep Kit for NeoPrep. Libraries were normalized to 10 nM by the Neoprep.
Each library quantity was checked again using Qubit DNA HS Assay Kit and 2 l of each library were pooled together into a single tube. This pool of 16 samples was checked on the Agilent High Sensitivity D1000 ScreenTape to ensure the libraries were the correct size at 300 bp. The pooled sample was diluted to 4 nM. The 4 nM library (containing the 16 pooled libraries) was sequenced on a NextSeq500 using a NextSeq® 500/550 High Output Kit v2 (150 cycles) with a 1% PhiX control spiked in.
Counts per gene were calculated using the htseq-count tool from the HTSeq v0.6.1p1 package (RRID:SCR_005514) (4) with the following parameters: --format=bam --minaqual=10 --stranded=reverse --mode=union. Differentially expressed genes were identified using the DESeq2 v1.14.1 package (RRID:SCR_015687) (5) from Bioconductor release 3.3. Differentially expressed genes were called at a false discovery rate of 5%. Adjusted p-values for the KD-shRNA vs SCR-shRNA pairwise comparison were re-calculated using fdrtool (6). Pathway Analysis was carried out using GAGE (7) v2.22 package from Bioconductor release 3.3, referencing KEGG pathways (RRID:SCR_012773) and assessing gene sets towards both single directions (up-/down-regulated) and both directions simultaneously (bi-directional). Differentially regulated pathways were called at a p value of <0.01. The accession number for the raw and processed data files for the RNA sequencing analysis reported in this paper is NCBI GEO: GSE106873.

Whole metabolome analysis
Metabolism was quenched in a) KD siRNA and SCR siRNA cells growing in 6-well plates, 96 hours after cell loading and 72 hours after siRNA transfection and b) KD shRNA and SCR shRNA cells, 96 hours after cell loading. Six biological replicates were used per cell group, each consisting of two wells. 200 l of media from each well were collected and frozen, followed by media removal and washes with three aliquots of Phosphate-Buffered Saline. Following this, 0.9 ml of a 40/40/20% solution of acetonitrile, methanol and water (Sigma-Aldrich) at a temperature of approximately -40 o C was added to each well and plates were frozen at -80 o C for 15 min. Cells were scraped off the wells and the suspension was centrifuged at 12,000 rpm (4 o C) for 10 min to separate the extraction supernatant from cell pellet. The extraction supernatant was dried applying a vacuum centrifugal evaporator (Thermo Scientific Savant SPD111V speedvac concentrator coupled to a Savant RVT5105 vapour trap). 100 l of media samples were dried applying the same process. Pooled QC samples were prepared for cell media analysis (by pooling of 100 l aliquots of all biological samples) and cell extract analysis (by pooling of 200 l aliquots of all cell extract samples).
Samples were analysed applying ultra-performance liquid chromatography-mass spectrometry (UPLC-MS) using a Thermo Scientific Ultimate3000 UPLC system coupled to an electrospray Q Exactive Focus mass spectrometer. Two assays were applied to increase the number of metabolites detected, a HILIC assay to investigate water-soluble metabolites and a C18 reversed phase method to investigate lipid metabolites.
The HILIC method applied a Thermo Scientific Accucore 150 Amide HILIC column (100 x 2.1mm, 2.6 µm) operated at a temperature of 35°C and a flow rate of 500 µL.min -1 . Solvent A was A 10 mM Ammonium Formate in 95% Acetonitrile/5% water + 0.1% formic acid and solvent B was 10 mM Ammonium Formate in 50% Acetonitrile/50% water + 0.1% formic acid. The gradient elution was applied at follows; Start at 99%A for 1 minute, followed by decreases to 85% and 50% at 3 minutes and 6 minutes with a curve of 5 and then a decrease to 5% A at 9 minutes and an increase to 99% A at 10.5 minutes with a curve of 5. The total analysis time was 15 minutes and the injection volume was 2µL.
Mass spectral data was collected in positive and negative ion modes separately at a mass resolution of 70,000 (FWHM at m/z 200). Data Dependent Analysis data was acquired for three QC samples to aid metabolite identification. QC samples were analysed 10 times at the start of the run and then after every 6 th biological sample with two QC samples analysed after all biological samples had been analysed.
The C18 reversed phase method applied a Thermo Scientific Hypersil GOLD C18 column (100 x 2.1mm, 1.9 µm) operated at a temperature of 55°C and a flow rate of 400 µL.min -1 . Solvent A was 10 mM Ammonium Formate in 60% Acetonitrile/40% water + 0.1% formic acid and solvent B was 10 mM Ammonium Formate in 90% isopropyl alcohol/10% acetonitrile + 0.1% formic acid. The gradient elution was applied at follows; Start at 80%A for 0.5 minutes, followed by a decrease to 0% A at 8.5 minutes with a curve of 5 and then an increase to 80% A at 11.5 minutes with a curve of 5. The total analysis time was 15 minutes and the injection volume was 2µL. Mass spectral data was collected in positive and negative ion modes separately at a mass resolution of 70,000 (FWHM at m/z 200). QC samples were analysed 10 times at the start of the run and then after every 6 th biological sample with two QC samples analysed after all biological samples had been analysed.
Raw data was converted to the mzML format applying ProteoWizard and then deconvoluted applying the software XCMS (9), operated on an office PC in R using previously described parameters (10). Putative metabolite annotations were provided using the software PUTMEDID_LCMS using a RT window of +/-2 seconds and a mass error of 5ppm (11).
The data were filtered for quality based on the QC sample data with metabolites with a relative standard deviation >20% or detected in less than 60% of the QC samples being removed (12).
Univariate and multivariate data analysis and pathway enrichment analysis were performed in MetaboAnalyst 3.0 (RRID:SCR_015539) (13). This included Principal Components Analysis (PCA), Mann Whitney U tests or Kruskal-Wallis tests to identify metabolites demonstrating a statistically significant change in relative concentrations between two or three biological classes. Fold changes were calculated by division of the mean peak response for one biological class by the mean peak response of the second biological class.