OCR Analysis

HepG2 cells were collected by trypsinization and seeded on a collagen-coated XF 96-well cell culture microplate (25 000 cells/100 µl medium/well) and incubated overnight in glucose medium (37 °C, 5% CO₂). All assays used ≤ 0.5% DMSO as a vehicle control.

Cells were incubated for 1 h (37°C, 0% CO₂) before culture medium was replaced by 175 µl of unbuffered Seahorse XF Base medium supplemented with glucose (25 mM), L-glutamine (2 mM), sodium pyruvate (1 mM), pre-warmed to 37 °C (pH 7.4). Prior to measurement of OCR, the Seahorse XFe96 instrument gently mixed the assay medium in each well for 10 min to enable the oxygen partial pressure to reach equilibrium. The OCR was then measured 3 times to establish a baseline rate prior to the acute injection of flutamide, 2-hydroxyflutamide or bicalutamide (7.8–500 µM). There were 9 OCR measurement cycles following compound injection and each measurement cycle consisted of a 3 min mix and 3 min measure. Following this compound incubation (54 min), a mitochondrial stress test was performed consisting of sequential injections of oligomycin (1 µM), carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP) (0.5 µM) and antimycin A/rotenone (1 µM each) (all compound concentrations were optimized to generate the maximum effect in the absence of toxicity). After stress test compound injections there were 3 measurement cycles before the injection of the next stress test compound. This enabled the calculation of basal respiration (OCR prior to oligomycin injection—non-mitochondrial OCR), proton leak (OCR after oligomycin injection—non-mitochondrial OCR), ATP-linked OCR (basal respiration—proton leak—non-mitochondrial OCR), maximal respiration (first injection after FCCP injection—non-mitochondrial OCR) and spare respiratory capacity (maximal respiration—basal respiration) (Figure 2).

Representative control mitochondrial stress test trace. Mitochondrial stress test assays consisted of a series of compound injections into the cell culture microplate. Flutamide/2-hydroxyflutamide or vehicle control (shown) was first injected, followed by 9 measurement cycles. Remaining injections consisted of oligomycin (ATP synthase inhibitor), FCCP (OXPHOS uncoupler), and rotenone/antimycin A (complex I and III inhibitors, respectively) with each followed by 3 measurement cycles. This series of manipulations enabled the calculation of parameters: basal, ATP-linked, maximum, and non-mitochondrial OCR, as well as proton leak. Each measurement cycle was a total of 6 min.

Culture medium was replaced with mitochondrial assay solution (MAS) buffer (MgCl₂; 5 mM, mannitol; 220 mM, sucrose; 70 mM, KH₂PO₄; 10 mM, HEPES; 2 mM, EGTA; 1 mM; BSA; 0.4% w/v) and plasma membrane permeabilizer (PMP) (1 nM) containing constituents to uncouple cells and stimulate oxygen consumption via complex I (ADP; 4.6 mM, malic acid; 30 mM, glutamic acid; 22 mM, BSA; 30 µM, PMP; 1 nM, FCCP; 8 µM) (All compound concentrations were optimized to generate the maximum effect in the absence of toxicity) and flutamide or 2-hydroxyflutamide (10–250 µM). PMP is a recombinant form of perfringolysin O, a cholesterol-specific pore-forming reagent which requires a higher threshold level of cholesterol than native perfringolysin O. This enables selective permeabilization of the cell membrane whilst having little or no effect on cholesterol-deficient mitochondrial membranes (Divakaruni et al., 2013; Salabei et al., 2014). Permeabilization of the cell membrane not only permitted mitochondrial access to all of the substrates and inhibitors used, but also depleted the cells of all cytosolic stores, ensuring that mitochondrial electron transport could be specifically driven by the substrates provided. The use of FCCP-treated (uncoupled) cells ensured that any deviations seen were due to perturbations at specific respiratory complexes and not due to inefficiencies in the coupling of OXPHOS.

Following a basal measurement of 3 cycles of mix (30 s), wait (30 s), and measure (2 min), sequential injections of A: rotenone (2 µM), B: succinate + rotenone (20 mM, 2 µM, respectively), C: antimycin A (2 µM) and D: ascorbic acid + N,N,N′,N′-tetramethyl-p-phenylenediamine (TMPD) + antimycin A (20 mM, 0.5 mM, and 2 µM, respectively) were performed with a 2 cycle interval between each, allowing measurement of changes in complexes I (A), II (B), and IV (C and D) activity respectively (Figure 3). MAS buffer, all constituents and compound injections were used at pH 7.2.

Representative in situ respiratory complex assay trace. In situ respiratory complex assays consisted of cells in a solution containing substrates for complex I and flutamide/2-hydroxyflutamide or vehicle control (shown) prior to 3 cycles of measurements and a series of compound injections into the cell culture microplate. Injections consisted of rotenone (complex I inhibitor), succinate (complex II substrate), antimycin A (complex III inhibitor), and TMPD/ascorbate (complex IV substrates) with 2 cycles of measurements following each. This series of manipulations enabled the calculation of complex I (A), II (B), and IV (C) activity. Each measurement cycle was a total of 3 min.

Culture medium was replaced with MAS buffer containing constituents to stimulate oxygen consumption via complex I (as previously without FCCP), complex II (ADP; 4.6 mM, succinate; 20 mM, rotenone; 1 µM, BSA; 0.2% w/v, PMP; 1 nM), or complex III (ADP; 4.6 mM, duroquinol; 500 μM, rotenone; 1 μM, malonic acid; 40 μM, BSA; 0.2% w/v, PMP; 1 nM) dependent on the respiratory complex of interest. Following a basal OCR measurement of 3 cycles of mix (30 s), wait (30 s), and measure (2 min), flutamide/2-hydroxyflutamide were injected (10–250 µM) and 3 cycles of measurement made again, prior to a mitochondrial stress test as detailed previously but with changes to stress test compound concentrations; oligomycin (1 μM), FCCP (10 μM), rotenone/antimycin A (2 μM). Changes in complex II activity were also assessed at lower compound concentrations; 2–30 μM (Supplementary Figure S1). Complex I, II, and III activity were defined by the change in complex I, II, or III-stimulated maximal respiration respectively compared with vehicle control.

Culture medium was replaced with MAS buffer containing constituents to stimulate oxygen consumption via complex IV as this was not significantly affected by either compound in the in situ respiratory complex assay (ADP; 4.6 mM, ascorbic acid; 20 mM, TMPD; 0.5 mM, antimycin A; 2 µM, BSA; 30 µM, PMP; 1 nM). The assay consisted of a basal OCR measurement of 2 cycles of mix (30 s), wait (30 s), and measure (2 min) followed by MAS or FCCP injection (0.5 µM) and 2 measurement cycles. MAS-injected cells remain coupled whereas FCCP-injected cells become uncoupled meaning Complex V (ATP synthase) inhibition should not result in a change in OCR. Either flutamide, 2-hydroxyflutamide (10–250 µM) or oligomycin (positive control; 1 µM) was then injected into both the uncoupled and coupled cells, followed by a final 2 measurement cycles (Figure 4). Change in complex V activity was defined as the difference in flutamide/2-hydroxyflutamide-induced OCR change between coupled (MAS injection) and uncoupled (FCCP injection) cells compared with vehicle control.

Representative complex V assay trace. Complex V assays consisted of cells in a solution containing substrates for complex IV before a series of compound injections into the cell culture microplate. FCCP (OXPHOS uncoupler) (Trace A) or MAS buffer (Traces B and C) was first injected, followed by 2 cycles of measurements. Flutamide, 2-hydroxyflutamide, oligomycin (positive control), or vehicle control was then injected into both the uncoupled (FCCP-treated) and coupled (MAS-treated) cells. Change in complex V activity was defined as the reduction in OCR of coupled cells upon flutamide/2-hydroxyflutamide injection minus the change in OCR of uncoupled cells, as a % of vehicle control. Injections for traces shown; A, FCCP, 250 μM 2-hydroxyflutamide; B, MAS, 250 μM 2-hydroxyflutamide; C, MAS, Oligomycin. Each measurement cycle was a total of 3 min.

Following completion of OCR assays, medium was removed from all wells and 20 µl of somatic cell ATP releasing agent was added to each well before shaking for 1 min and transfer of 10 µl of lysate to a 96-well plate and performance of a BCA assay according to the manufacturer’s instructions. Protein content per well was used to normalize OCR.