ENDOCRINE SOCIETY
OXFORD
Mitotane Targets Lipid Droplets to Induce Lipolysis in Adrenocortical Carcinoma
Kate M. Warde,1.[D Yi Jan Lim,1 Eduardo Ribes Martinez,1 Felix Beuschlein,2,3,[D Paula O’Shea,4 Constanze Hantel,2,5 and Michael Conall Dennedy1,[D
1Discipline of Pharmacology and Therapeutics, National University of Ireland, Galway, H91 TK33, Ireland
2Department of Medicine IV, University Hospital, Ludwig Maximilian University of Munich, Munich, 81377, Germany
3Department of Endocrinology, Diabetes, and Clinical Nutrition, University Hospital Zurich, Zurich 8091, Switzerland
4Department of Clinical Biochemistry, Galway University Hospitals, Saolta Hospitals Group, Newcastle Road, Galway, H91 RW28, Ireland; and 5Medizinische Klinik und Poliklinik III, University Hospital Carl Gustav Carus Dresden, 01307, Germany
Correspondence: Michael Conall Dennedy, MD, PHD, FRCPI, School of Medicine, National University of Ireland, Galway, Costello Road, Galway, Ireland. Email: michael.dennedy@nuigalway.ie.
Abstract
Introduction: Adrenocortical carcinoma (ACC) is a rare aggressive cancer with low overall survival. Adjuvant mitotane improves survival but is limited by poor response rates and resistance. Mitotane’s efficacy is attributed to the accumulation of toxic free cholesterol, predominantly through cholesterol storage inhibition. However, targeting this pathway has proven unsuccessful. We hypothesize that mitotane-induced free- cholesterol accumulation is also mediated through enhanced breakdown of lipid droplets.
Methodology: ATCC-H295R (mitotane-sensitive) and MUC-1 (mitotane-resistant) ACC cells were evaluated for lipid content using specific BODIPY dyes. Protein expression was evaluated by immunoblotting and flow cytometry. Cell viability was measured by quantifying propidium iodide-positive cells following mitotane treatment and pharmacological inhibitors of lipolysis.
Results: H295R and MUC-1 cells demonstrated similar neutral lipid droplet numbers at baseline. However, evaluation of lipid machinery demon- strated distinct profiles in each model. Analysis of intracellular lipid droplet content showed H295R cells preferentially store cholesteryl esters, whereas MUC-1 cells store triacylglycerol. Decreased lipid droplets were associated with increased lipolysis in H295R and in MUC-1 at toxic mitotane concentrations. Pharmacological inhibition of lipolysis attenuated mitotane-induced toxicity in both models.
Conclusion: We highlight that lipid droplet breakdown and activation of lipolysis represent a putative additional mechanism for mitotane- induced cytotoxicity in ACC. Further understanding of cholesterol and lipids in ACC offers potential novel therapeutic exploitation, especially in mitotane-resistant disease.
Key Words: adrenocortical carcinoma, mitotane, lipid droplets, lipolysis, cholesterol, tumor resistance
Abbreviations: ACC, adrenocortical carcinoma; ATGL, adipose triglyceride lipase; CE, cholesteryl ester; FC, free cholesterol; HDL, high-density lipoprotein; HSL, hormone-sensitive lipase; LD, lipid droplet; LDL, low-density lipoprotein; pHSL, phosphorylated hormone-sensitive lipase; PI, propidium iodide; PLIN, perilipin; RT-qPCR, real-time quantitative PCR; SOAT1, sterol-O-acyl-transferase 1; TAG, triacylglycerol; tHSL, total hormone-sensitive lipase; VC, vehicle control
Adrenocortical carcinoma (ACC) is a rare aggressive cancer of the adrenal cortex, carrying a poor prognosis and a me- dian overall survival of 24 months in advanced disease (1). Current first-line therapy is complete (R0) surgical resection (2, 3). However, local and distant metastatic recurrence is fre- quent and, consequently, adjuvant chemotherapy with the adrenocorticolytic agent mitotane has become the standard of care for most patients (4, 5). Mitotane is currently the only licensed pharmacological therapy for ACC that im- proves recurrence-free survival (5). However, it is limited by its narrow therapeutic window, toxic adverse effects, and resistance to therapy (6-8). Mitotane’s mechanism of action is poorly understood. It has been shown in several studies to exert its adrenocorticolytic effect at least in part through lipotoxicity induced by intracellular free cholesterol (FC) ac- cumulation (9-11). This has been previously attributed to the inhibition of cholesterol storage primarily through action on sterol O-acyltransferase 1 (SOAT1) (9).
The steroidogenicity of the adrenal cortex dictates the requirement for a constant supply of precursor cholesterol (12-14). Adrenocortical cholesterol is sourced through 4 potential processes: (1) intracellular de novo synthesis; (2) uptake of circulating high-density lipoprotein (HDL) via scavenger receptors; (3) uptake of circulating low-density lipoprotein (LDL) via receptor-mediated endocytosis; and finally (4) mobilization of stored intracellular cholesteryl esters (CEs), converted to FC by hormone-sensitive lipase (HSL) and liberated from lipid droplets (LDs) (12, 15-18). Disruption of any of these pathways, through genetic or pharmacological manipulation, has the potential to signifi- cantly affect adrenal function and morphology. At rest, chol- esterol metabolism in adrenocortical cells is predominantly focused on intracellular storage and efflux (12). However, acute stress responses rapidly mobilize cholesterol from intracellular LD stores to the mitochondria for steroid syn- thesis (13).
In adipose tissue and liver, LDs were traditionally con- sidered inert storage organelles. However, greater under- standing of these organelles demonstrates that they actively mediate intracellular lipid and cholesterol flux, dynamic- ally responding to intracellular need in supporting meta- bolic and energy requirements (19-21). The dynamic role of LDs is largely mediated by LD-associated proteins, namely perilipins (PLINs), CIDEC, and CGI-58 (21). These proteins dynamically interact with lipids, CEs, triacylglycerol (TAG), cytoplasmic proteins, and enzymes such as HSL and adipose triglyceride lipase (ATGL) to modulate flux into and out of the LD, as well as lipid/cholesterol storage within (22-24).
Adrenocortical lipid storage and LDs remain largely un- explored but are regarded as predominantly CE-storing or- ganelles (25, 26). This assumption is broadly determined by the essential requirement of intracellular cholesterol for steroidogenesis. Adrenocortical lipolysis is primarily medi- ated by HSL and ATGL for CEs and TAG, respectively (14, 20). The role of HSL in adrenocortical cells as a neutral chol- esterol ester hydrolase, as well as its importance in facilitating steroidogenesis have been previously described (15, 20, 24). Critically, in the context of cancer, LDs have been identified as key mediators of oncogenesis mediating essential roles in cancer cell survival and metastasis (27-30).
The complexity of lipid, lipoprotein, and cholesterol inter- actions in the context of ACC and mitotane therapy has previously been demonstrated and speaks to the intricate pharmacodynamics of mitotane (10, 31-33). In ACC, excess circulating lipids are a common side effect of mitotane and HMGCR inhibitors (statins) are often used as a method of managing hypercholesterolemia (34). Additionally, the use of statins has also demonstrated a significant increase in disease control in a small cohort of ACC patients (10). Although this effect is supported by in vitro work, coadministration of statins and mitotane has limited overall effectiveness, possibly because of opposing effects of each respective drug on intra- cellular lipid levels (9, 35). There clearly exists a significant need for further detailed understanding of adrenocortical lipid, cholesterol, and lipoprotein metabolism in the context of mitotane therapy to aid in identifying new adjunct thera- peutics to potentiate or mimic mitotane effect and efficacy.
In the current study, we focus on the interaction between mitotane and the lipid droplet. We highlight novel differ- ences in LD storage dynamics between the mitotane-sensitive H295R model and mitotane-resistant MUC-1 model of ACC. In particular, H295R demonstrates a CE-rich phenotype, whereas the phenotype of MUC-1 is CE-poor and relatively triacylglycerol-rich. We hypothesize mitotane’s effect on CE breakdown through hormone sensitive lipase as a poten- tial mechanism of effective mitotane-induced lipotoxicity in H295R. Although MUC-1 demonstrates resistance to these effects within the tolerated therapeutic range, we confirm this mechanism of action at the supratherapeutic concentrations of mitotane necessary to induce cell death within this cell line.
Materials and Methods
Cell Culture and Treatments
H295R human primary ACC cells (NCI-H295R, CRL-2182 [American Type Culture Collection, RRID:CVCL_0458]), MUC-1 human metastatic ACC cells, and HepG2 human liver cancer cells (HB-8065 [American Type Culture Collection,
RRID:CVCL_0027]) were cultured and validated as previ- ously described (33, 36).
Cells were treated at indicated concentrations using CAY10499: HSL inhibitor (Cayman Chemical), Atglistatin: ATGL inhibitor and mitotane (2,4’-DDD). All drugs were prepared in dimethyl sulfoxide and vehicle control remained below 0.1%. Cells were pretreated with CAY10499 (20 uM) for 18 hours followed by mitotane (at indicated concentra- tions) for 6 hours. Cells were cholesterol-loaded using 45 µg/ mL water-soluble cholesterol methyl-ß-cyclodextrin 1 hour before mitotane treatment. Reagents were obtained from Sigma unless otherwise indicated.
Gene Expression
Following drug treatments, cells were washed with cold PBS, and total RNA was extracted from adherent cells as previously described (33). Real-time quantitative PCR (RT-qPCR) expression was assayed on Applied Biosystems StepOne Plus (Applied Biosystems) and carried out using SYBR green mastermix (Promega) using the following pri- mers (5’-3’); LIPE [FWD: CTATGCTGGTGCAAAGAC; REV: CTCCAGGAAGGAGTTGAG], PLIN2:
[FWD: ATGGCATCCGTTGCAGTTGAT; REV:
GGACATGAGGTCATACGTGGAG],
PLIN3:
[FWD: TATGCCTCCACCAAGGAGAG;
REV:
ATTCGCTGGCTGATGCAATCT], ACTB [FWD:
CACCATTGGCAATGAGCGGTTC; REV:
AGGTCTTTGCGGATGTCCACGT], B2M [FWD:
CCACTGAAAAAGATGAGTATGCCT; REV: CCAATCCAAATGCGGCATCTTCA]. Fold change was cal- culated using the 44CT method. Data are represented as fold change of mRNA relative to vehicle control. Reference gene B2M was chosen as the optimal comparator between H295R and MUC-1 gene expression using Qbase+ selection tool.
Protein Expression
Samples were harvested on ice and lysed using radioimmunoprecipitation buffer and 1X protease inhibitor cocktail (Calbiochem). Lysates were quantified, separated by SDS-PAGE and transferred to polyvinylidene fluoride membrane as previously described (33). Membranes were incubated with primary antibodies overnight followed by relevant horseradish peroxidase-linked secondary antibodies: pHSLser563 (CST #4139: 1:2000), RRID:AB_2135495), pHSLser660 (CST #45804: 1:2000, RRID:AB_2893315); HSL (SCBT #74489 1:1000, RRID:AB_2135504); PLIN1 (Biotechne AF6615: 1:1000, RRID:AB_10717859); PLIN2 (Biotechne, MAB7634 1:2000 RRID:AB_2920789); PLIN3 (SCBT #390981 1:1000, RRID:AB_2920790); PLIN4 (Progen #GP34S 1:500, RRID:AB_2920687); SOAT1 (SCBT #69836 1:1000, RRID:AB_1129582); DGAT1 (SCBT #271934 1:500, RRID:AB_10649947); and ß actin (Sigma; A5316; 1:5000, RRID:AB_476743). Imaging was carried out as pre- viously described (33).
Cell surface expression of lipid uptake receptors was car- ried out using flow cytometry. Cells were harvested via trypsinization and washed in 1X cold PBS. Staining was car- ried out using the following antibodies according to manufac- turers’ guidelines for 30 minutes in the dark at 4℃; SCARB1/ CD36L1 (PE; Biolegend #363203, RRID:AB_2564208), LDLR (PE; Becton Dickinson #565653, RRID:AB_2739325), LRP1/CD91 (APC; Miltenyi Biotec 130-111-414,
RRID:AB_2659600), and CD36 (FITC; Miltenyi Biotec #130-110-876, RRID:AB_2657726). A total of 10 000 events were acquired using the FACS Canto II (BD Bioscience) and analysis was carried out using FlowJo V10 analysis software (Tree Star Inc, RRID:SCR_008520).
Cell Death
For quantitative cell death analyses, cells were trypsinized and resuspended in FACS buffer (PBS, 1% fetal bovine serum, 0.05% Sodium Azide) following treatment and stained for 5 minutes using propidium iodide (PI) (5 µg/mL). PI was ex- cited using a 488-nm laser followed by 585/35 bandpass filter. Flow cytometry was performed on FACS CantoII with FACS DiVa 6.0 acquisition software (BD Biosciences) and FlowJo V10 analysis software (Tree Star Inc). Gating strategy is as previously described (33).
Lipid Analysis
BODIPY 493/503 (Thermofisher, D3922) was used to stain neutral and nonpolar lipids (37). Cells were harvested via trypsinization, and 1 × 106 cells stained using 5 uM BODIPY 493/503 for 30 minutes at 4℃ in the dark. Cells were washed and resuspended in 100 uL of FACS buffer. Before acquisition, 5 µg/mL of PI was added per sample for 5 minutes and 10 000 events were recorded. Fluorescence was detected using a 488- nm excitation laser followed by a 530/30 (BODIPY 493/503) and 585/355 (PI) bandpass filters.
BODIPY FL C12 cholesterol ester; cholesteryl 4,4- difluoro- 5,7-dimethyl-4-bora-3a,4a-diaza-s-indacene-3-dodecanoate (Thermofisher, C3927MP) is a green fluorescent choles- terol ester analogue. BODIPY 558/568 C12; 4,4-difluoro-5- (2-thienyl)-4-bora-3a,4a-diaza-s-indacene-3-dodecanoic acid (Thermofisher, D3835) is an orange-red fluorescent fatty acid dye. Each identifies cholesteryl ester and triglyceride- containing lipid droplets, respectively (22). Cells were grown overnight in media supplemented with 1 uM BODIPY FL C12 cholesterol ester (hereafter referred to as CE-BODIPY) and 5 uM BODIPY 558/568 C12 (referred to as TAG-BODIPY). Cells were harvested via trypsinization and resuspended in FACS buffer. A total of 10 000 events were acquired using the FACS Canto II (BD Bioscience). Sytox Blue (Invitrogen #S34857) was used as a viability dye. Fluorescence was re- corded using a 488-nm excitation laser followed by a 530/30 bandpass filter (CE-BODIPY) or a 585/35 bandpass filter (TAG-BODIPY) and 405-nm excitation laser followed by a 450/50 bandpass filter (Sytox Blue).
Additionally, cells were analyzed using the same methodology via imaging flow cytometry using the Amnis ImageStream™Mk II (Luminex Corporation). High sensitivity mode (60X) was used to record 5000 events/sample. The population of live, single, focused cells was used for analysis of BODIPY493/503 or CE-BODIPY according to Intensity_MC_Ch02 and TAG- BODIPY according to Intensity_MC_Ch03. Image display properties were standardized and uniformly applied to each .daf; live cells were gated and resolution metric was calculated as described (33). Lipid droplet number (Spot Count Score) was calculated using the spot count tool in IDEAS as previ- ously described (33). Data are represented as number of fluor- escence spots (lipid droplets) per individual cell.
Cholesteryl esters were quantified using the Fluorometric Cholesterol Assay Kit (Cayman Chemical, #10007640) ac- cording to manufacturer guidelines. Briefly, H295R cells were
treated with indicated concentrations of mitotane for 6 hours. Following treatment, cells were washed with ice cold 1X PBS and incubated with hexane:isopropanol (3:2) for 1 hour and placed on a rocker. The supernatant was transferred to glass vials and dried at room temperature. Remaining lipids were resuspended and quantified according to the kit guidelines following a 1:5 dilution.
Glycerol Quantification
Glycerol was measured using a the MAK117 kit (Sigma) ac- cording to manufacturer guidelines. Glycerol concentrations were interpolated from known standards and adjusted for background using media only samples. Protein standardiza- tion was carried out by lysing the adherent fraction in each well using 10 uL radioimmunoprecipitation buffer and quan- tified by BCA assay.
Data Analysis
Statistical analyses were performed using GraphPad Prism V 8.0. Data are represented as means ± SEM unless stated other- wise. Paired sample analyses were performed using a 2-tailed Student t test. Multiple-group comparisons were carried out using an ANOVA followed by Tukey post hoc test. Statistical significance for 2-tailed analyses (P value) was assigned for values < 0.05.
Results
Mitotane-sensitive and Mitotane-resistant Cells Differ in their Lipid Storage
We used H295R and MUC-1 adrenocortical carcinoma cell lines as models of mitotane-sensitive and -resistant disease, respectively (33, 36). To compare lipid storage in these model cell lines, we stained lipids with a neutral lipid fluorescent dye, BODIPY 493/503, and quantified the amount of staining by conventional cytometry and imaging flow cytometry. Both cell lines contained abundant neutral lipids within droplets as measured by flow cytometry and seen in representative microscopy images (Fig. 1A-D). To compare fluorescence be- tween both cell lines, a resolution metric was calculated that normalizes differences in autofluorescence. Although MUC-1 cells stained more than H295R cells, this was not a statistic- ally significant difference (H295R, 0.9824 vs MUC-1, 1.679 [P = 0.0525]) (Fig. 1E). The number of fluorescent spots in the cells, indicative of the number of lipid droplets they contain, was similar for both cell lines (H295R, 2.29 vs MUC-1, 2.31 [P = 0.987]) (Fig. 1F).
Lipid droplets store predominantly either cholesteryl ester or triacylglycerol (22, 23). Because overall lipid droplet num- bers were similar in both cell lines, we hypothesized that dif- ferences in CEs and triacylglycerol may explain the difference observed in mitotane sensitivity. To quantify and visualize CE and triacylglycerol lipid droplets in H295R and MUC-1 cells, we stained them with BODIPY FL C12 CE (CE-BODIPY) and BODIPY 588/568 C12 (TAG-BODIPY; as a marker for fatty acids in triacylglycerol) and analyzed them (as previously) by flow cytometry and imaging flow cytometry. H295R cells were rich in Ces, whereas MUC-1 cells were relatively CE-poor (Fig. 1G). Triacylglycerol fluorescence was similar in H295R and MUC-1 cells in the absence of cholesterol and the presence of cholesteryl increased triacylglycerol staining to similar levels in both cell lines (Fig. 1H). Because MUC-1 cells stained
A
H295R
B
BODIPY 493/503
Overlay
Brightfield
C
MUC1
D
Ch02
Ch01/Ch02
Ch01
BODIPY 493/503
Overlay
Brightfield
2.5
2.5
Ch02
Ch01/Ch02
Ch01
UNSTAINED
Normalized Frequency
UNSTAINED
2
Normalized Frequency
2
STAINED
1.5
1.5
STAINED
1
1.
0.5
0.5
0
0
1e3
1e4
1e5
1e6
1e7
==== 7 um
1e3
1e4
1e5
1e6
1e7
BODIPY 493/503
BODIPY 493/503
… 7 pm
E
BODIPY 493/503
F
BODIPY 493/503
2.0
p=0.0525
p=0.9870
3
G
Fluorescence Intensity (Resolution Metric)
i. H295R Baseline
CE-BODIPY
FA-BODIPY
OVERLAY
BRIGHTFIELD
1.5-
2858
Spot Count
2-
1.0-
***. 7 um
0.5-
1-
ii. H295R Cholesterol
CE-BODIPY
FA-BODIPY
OVERLAY
BRIGHTFIELD
0.0
0
539
H295R
MUC-1
H295R
MUC-1
**** 7 um
H
CE-BODIPY
TAG-BODIPY
iii. MUC-1 Baseline
CE-BODIPY
FA-BODIPY
OVERLAY
BRIGHTFIELD
2.0
p=1.76x10-2
H295R
902
Fluorescence Intensity (Resolution Metric)
MUC-1
p=2.70x10-3
1.5
p=0.5287
… 7 um
iv. MUC-1 Cholesterol
1.0
p=0.5531
CE-BODIPY
FA-BODIPY
OVERLAY
BRIGHTFIELD
2177
T
0.5
… 7 um
0.0
T
T
I
I
T
1
Cholesterol
-
+
-
+
-
+
-
+
45 µg/ml
relatively lower for CEs, we concluded that triacylglycerol is likely the main component of LDs in these cells. Additionally, we cultured each cell line in the presence of water-soluble chol- esterol to assess the ability of each cell line to take up and store cholesterol. In H295R, we saw a 1.7-fold increase in CE fluor- escence; however, in MUC-1 cells, there was no significant difference in CE fluorescence upon culture in the presence of cholesterol, indicating that MUC-1 cells show lower ability to take up or store cholesterol (Fig. 1G). Thus, mitotane-sensitive and -resistant cells differ in the main lipid component of their lipid droplets: the sensitive cells store predominantly CEs and the resistant cells predominantly triacylglycerol.
Mitotane-sensitive and Mitotane-resistant Cells Differ in Their Lipid Handling Proteins
Because the lipid droplets in mitotane-sensitive cells contain predominantly CEs and those in mitotane-resistant cells contain
predominantly triacylglycerol, we next investigated whether they also differ in the proteins associated with them by immuno- blotting H295R and MUC-1 cells with antibodies specific for perilipins 1 through 4 (PLIN1-4). The lipid-rich liver cancer cell line HEPG2 served as a positive control. H295R cells contained little or no PLIN1 but were relatively rich in PLIN2, whereas MUC-1 cells contained relatively small amounts of both pro- teins; the amounts of PLIN3 and PLIN4 were similar in both cell lines (Fig. 2A). In agreement with the protein levels meas- ured by immunoblotting, PLIN2 mRNA expression was lower in MUC-1 than in H295R cells as determined by RT-qPCR and normalization to the reference mRNA B2M (H295R vs MUC- 1, -2.8-fold [P = 0.0019]) and there was no difference between relative PLIN3 mRNA levels in the 2 cell lines (H295R vs MUC-1, 0.4-fold [P = 0.5241]) (Fig. 2B and 2C).
To compare the relative levels of key lipid-uptake recep- tors on the surface of H295R and MUC-1 cells, we labeled
the cells with fluorescent antibodies and quantified them by flow cytometry. The expression profiles for lipid-uptake re- ceptors differed significantly between H295R and MUC-1. Most notably, the HDL uptake receptor (SCARB1) and the chylomicron remnant receptor (LRP1) were present in much larger amounts on H295R than on MUC-1 cells (H295R, 901.2 vs MUC-1, 32.4 [P < 0.0001] for SCARB1; H295R, 581.0 vs MUC-1, 135.6 [P < 0.0001] for LRP1) (Fig. 2D). Because cholesterol is the primary substrate for both SCARB1 and LRP1 (38, 39), this suggests that the H295R cell line may use cholesterol more than MUC-1 cells. Cell surface expres- sion of CD36 and LDLR were significantly lower than those of SCARB1 and LRP1, as indicated by flow cytometry (Fig. 2D). The fluorescence intensity of CD36, which functions pri- marily as a scavenger of modified LDL and a fatty acid uptake receptor (40), was 12-fold greater in H295R than in MUC-1 cells (H295R, 3.67 vs MUC-1, 0.299 [P < 0.0001]), pos- sibly indicating different metabolic requirements of the 2 cell lines, whereas LDLR levels were similar on both MUC-1 and H295R cells (H295R, 1.163 vs MUC-1, 1.823 [P = 0.6882]).
Hormone-sensitive lipase is a key enzyme in adrenocortical lipolysis. When we analyzed the amounts of this protein in the 2 cell lines by immunoblotting, we found more than 100- fold more HSL protein in H295R cells than in MUC-1 cells. The relative fold change is represented numerically below the immunoblot as calculated using densitometry (Fig. 2E) (15). Likewise, expression of LIPE mRNA (which encodes the HSL protein) was nearly 100-fold lower in MUC-1 than in H295R cells as determined by RT-qPCR following normalization to B2M reference gene (Fig. 2F) (H295R vs MUC-1, 98-fold [P = 0.0407]). We also analyzed the amounts of SOAT1 and DGAT1, 2 enzymes involved in LD storage, by immunoblot- ting. SOAT1 was present in similar amounts in H295R and MUC-1 cells (Fig. 2G) and in the lipid-rich HEPG2 cells. By contrast, DGAT1 levels were higher in MUC-1 than in H295R cells (Fig. 2H) and were much lower than in HEPG2 cells.
From these data, we conclude that mitotane-sensitive H295R cells and mitotane-resistant MUC-1 cells differ in their levels of key proteins involved in cholesterol uptake and in the major types of lipids they store: the mitotane-sensitive cells store cholesterol esters, they are rich in the receptors that mediate cholesterol uptake and in the lipolysis enzyme hormone-sensitive lipase, whereas the mitotane-resistant cells store triacylglycerol rather than cholesterol esters and they are relatively poor in their expression of receptors and en- zymes that handle cholesterol.
Mitotane Depletes Lipid Stores and Increases Lipolysis
Given the differences in cholesterol uptake and storage in mitotane-sensitive cells when compared to mitotane-resistant cells, we investigated the effects of mitotane on the amount of cholesterol and triacylglycerol in the mitotane-resistant line, H295R. Treatment of cells for 6 hours with therapeutic doses of mitotane significantly reduced the number of neu- tral lipid droplets per cell when compared with cells treated with the vehicle alone (0 uM, 1 vs 20 uM, 0.61 [P = 0.0059]; 0 µM, 1 vs 40 uM, 0.55 [P = 0.0029]) determined, as be- fore, by imaging flow cytometry (Fig. 3A and 3B), indicating that this drug depletes lipid stores in mitotane-sensitive cells. We measured the effect of mitotane on cholesteryl es- ters by quantifying the fluorescence marker CE-BODIPY in
individual H295R cells by flow cytometry and found a sig- nificant decrease in cholesteryl esters at 3 doses (0 uM, 121 vs 20 µM, 86 [P = 0.0134]; 0 uM, 121 vs 40 uM, 89 [P = 0.0252]; 0 uM, 121 vs 50 uM, 69 [P = 0.001]) (Fig. 3C). This effect was confirmed by absolute quantification of cholesteryl esters in H295R cells and treatment with mitotane found a 2-fold reduction in response to 20 uM mitotane (Supplementary Figure 1A (41)). These doses of mitotane also induced a sig- nificant decrease in triacylglycerol-labeled lipid droplets as determined by quantifying FA-BODIPY (0 uM, 153 vs 20 µM, 9 [P = 0.0133]; 0 uM, 153 vs 40 uM, 77 [P = 0.0013]; 0 µM, 153 vs 50 uM, 91 [P = 0.01]) (Fig. 3D).
Previous studies found that mitotane increases the levels of intracellular FC (9, 10, 33), which is the breakdown product of CEs. To investigate whether it also increases production of glycerol, the breakdown product of triacylglycerol, we quan- tified glycerol in the medium of H295R cells after mitotane treatment using a colorimetric assay. Consistent with the decrease in triacylglycerol in these cells following mitotane treatment, we found an 8-fold increase in glycerol in the me- dium of H295R cells treated with 40 uM mitotane when compared with those treated with the vehicle control (0 uM, 0.0578 vs 40 uM, 0.47 mM glycerol/ug protein [P = 0.002]) (Fig. 3E), indicating that mitotane promotes the breakdown of triacylglycerol to glycerol as well as the breakdown of CEs to FC in mitotane-sensitive cells (33).
Because mitotane decreases the amount of CEs in lipid droplets (shown previously) and is reported to increase levels of intracellular FC (9, 10, 33), we hypothesized that the drug might induce expression of HSL, the enzyme that catalyzes the conversion of CEs to FC. To address this hypothesis, we used RT-PCR to measure the relative levels of the LIPE mRNA in H295R cells treated with mitotane. LIPE mRNA levels were up to 5-fold higher in the cells treated with 50 uM mitotane than in the control H295R (0 uM vs 20 uM: 2.7-fold [P = 0.1633]; 0 uM vs 40 uM: 4.8-fold [P = 0.0042]; 0 uM vs 50 µM: 4.8-fold [P = 0.0048]) (Fig. 3F). We also measured HSL protein levels by immunoblotting H295R cells treated with 50 uM mitotane. Total HSL (tHSL) and HSL phosphor- ylated (pHSL) on Ser563 (the active form of this enzyme), rapidly increased following addition of mitotane. The amount of pHSL continued to increase over the course of at least 2 hours (Fig. 3G). These increased levels of pHSL were accom- panied by a significant and dose-dependent, mitotane-induced decrease in the lipid droplet-associated proteins PLIN1 and PLIN3 (Fig. 3H and 3I). Thus, the loss of cholesteryl esters and increase in FC in mitotane-treated cells is accompanied by increased expression of the active lipolytic enzyme HSL and decreased expression of perilipins, which act as “gate- keepers” at the lipid droplet surface.
Inhibition of Lipolysis Attenuates Mitotane-induced H295R Cell Death
To investigate whether increased lipolysis might account for the therapeutic effect of mitotane in inducing the death of ACC cells by producing toxic lipid products, we treated H295R cells with CAY10499, a specific pharmacological in- hibitor of HSL (hereafter referred to as HSLi). As expected, HSLi significantly increased the amount of fluorescently la- belled cholesteryl ester and triacylglycerol lipid droplets in the H295R cells (Supplementary Figure 1B and 1C (41)); moreover, the amount of active, pHSL in cells treated with
A
H295R MUC-1 HEPG2
B
PLIN2
C
PLIN3
2-
p=1.90×103
Relative mRNA Expression
3.
PLIN1
Relative mRNA Expression
1-
2-
p=0.5241
0-
PLIN2
1.
-1-
PLIN3
-2-
0-
-3-
-1
PLIN4
-4
-2
H295R
MUC-1
H295R
MUC-1
ACTIN
Lipid Uptake Receptor Expression
D
p<1.00x104
p<1.00x104
E
F
LIPE
1000
T
p=4.04x10-2
Fluorescence Intensity Resolution Metric
800
H295R MUC-1 HEPG2
Relative mRNA Expression
2
H295R
600
MUC-1
HSL
0
400
B Actin
-50-
200
1
-137
-343 Fold change
p<1.00x104
-100-
5
p=0.6882
-150
0
CD36
LDLR
SCARB1
LRP1
H295R
MUC-1
G
H295R MUC-1 HEPG2
H
H295R MUC-1 HEPG2
SOAT1
DGAT1
B Actin
B Actin
1
-1.2
1.7
Fold change
1
12
46
Fold change
mitotane was attenuated in the presence of the inhibitor (Fig. 4A and Supplementary Figure 1D (41)). HSLi alone had no statistically significant effect on H295R cell viability as meas- ured by quantification of propidium iodide-positive dead cells by flow cytometry (Fig. 4B), however, it somewhat at- tenuated mitotane-induced cell death (mitotane alone, 78% vs mitotane + HSLi, 66% [P = 0.0003]) (Fig. 4B). These data indicate that lipolysis by HSL may contribute to mitotane- induced cell death. Moreover, the mitotane-induced deple- tion of cholesteryl esters in H295R cells, observed above, was fully prevented by HSLi (HSLi, 229 vs HSLi + mitotane, 239 [P = 0.9967]) (Fig. 4C), indicating that HSL is the primary lipolytic enzyme in these cells and that cholesteryl ester levels correlate with cell viability.
In contrast to its effects on mitotane-induced deple- tion of cholesterol esters, HSLi only partially prevented mitotane-induced depletion of triacylglycerol (HSLi, 254 vs HSLi + mitotane, 185 [P < 0.001]) (Fig. 4D). We investi- gated the possible relevance of triacylglycerol depletion for mitotane-induced cell death by using an inhibitor of adi- pose triglyceride lipase, atglistatin (hereafter referred to as ATGLi) (43). This inhibitor attenuated mitotane-induced H295R cell death to a similar extent as did HSLi (mitotane, 85% vs ATGLi + mitotane, 68% [P < 0.0038]), whereas alone it had no effect (Fig. 4E). This indicates that lipolysis of triacylglycerol may also play a role in mitotane-induced cell death. Treatment of cells with both ATGLi and HSLi did not attenuate mitotane-induced cell death more than either agent
A
B
H295R
Vehicle Control
Mitotane 20 UM
Mitotane 40 uM
Spot Count Score (normalised to vehicle)
1.5
p=2.90×10-3
p=5.90×103
1.0
E
0.5
0.0
… 7 um
BODIPY 493/503: Neutral Lipid Droplets
0
20
40
Mitotane IM
C
H295R
D
H295R
E
Glycerol
p=1.00×10-3
p=1.00x102
p=2.52×10-2
200
p=1.30×10-3
0.6-
p=3.00x10-3
Cholestryl Esters (MFI CE-BODIPY)
150-
p=1.34×10-2
p=1.33×102
Triacylglycrol (MFI FA-BODIPY)
150
Glycerol mM/ µg protein
100-
0.4-
100
50-
50
0.2-
0
0.0-
Mitotane (μM)
0-
0
20
40
50
Mitotane (uM)
0
20
40
50
Control Mitotane 40 μΜ
F
LIPE
G
Control
Mitotane 50μM
H
Mitotane (uM)
p=4.80×103
Relative mRNA Expression
6
30
60
120
240
(mins)
0
20
40
50
p=4.20×10-3
pHSL ser563
PLIN3
4
p=0.1633
tHSL
B Actin
B Actin
1
-1.2
-1.8
-5.0
Fold change
2
1
1.7
1.9
1.5
2.6
pHSL:HSL Fold change
Mitotane (µM)
0
0
20
40
50
0 20 40 50
PLIN1
Mitotane (UM)
B Actin
1
-1.9
-1.6
-8.5
Fold change
alone (Fig. 4E). Indeed, in the absence of mitotane, treatment of H295R cells with a combination of HSLi and ATGLi sub- stantially increased cell death when compared with untreated cells or with cells treated with either agent (control, 12% vs ATGLi + HSL, 45% [P < 0.0001]) (Fig. 4E). This indicates the importance of lipolysis for the normal function of these cells. Overall, inhibition of lipolysis by using either HSLi or ATGLi prevented mitotane-induced H295R cell death.
Mitotane Induces Lipolysis in MUC-1 Cells only at Supratherapeutic Concentrations
Previous studies found that the MUC-1 cell line is re- sistant to mitotane-induced toxicity at typical therapeutic
concentrations but is susceptible at supratherapeutic con- centrations (33). To investigate whether these high concen- trations of mitotane have similar effects on lipid droplet numbers in MUC-1 to those seen in H295R cells, we treated MUC-1 cells for 24 hours with 50 uM mitotane (the thera- peutic concentration) equivalent to 14.6 mg/L in patients (6), 100 or 200 uM (both supratherapeutic concentrations) la- beled neutral lipids with BODIPY 493/503 and used imaging flow cytometry to observed lipid droplets and conventional flow cytometry to quantify BODIPY 493/503 staining. Cells treated with 200 uM mitotane had significantly fewer neu- tral lipid droplets (0 uM, 1 vs 200 uM, 0.37 [P = 0.0151]) than had the cells treated with lower concentrations of the
A
B
H295R
Mitotane (µM)
100
p=3.00x10-4
0
0
40
40
HSLi
+
+
% Cell Death
80
-
-
pHSL
ser660
60
tHSL
40
p=0.9185
20
B Actin
0
1.0
1.0
6.7
2.7
pHSL:HSL Fold change
Mitotane (uM)
0
0
40
40
HSLi
-
+
-
+
C
H295R
D
H295R
E
H295R
p=5.00x10-3
Control
p=0.9967
p=5.70x103
100
p=3.80×10-3
Mitotane
Cholestryl Esters (MFI CE-BODIPY)
300-
p<1.00x10-
300-
p<1.00x104
T
T
Triacylglycrol (MFI FA-BODIPY)
Cell Death (%)
80
p<1.00x104
200-
200-
60
p<1.00x104
100-
100-
40-
p=0.9999
0-
20
Mitotane (uM)
0
0
40
40
0-
HSLi
Mitotane (uM)
0
0
40
40
-
+
-
+
HSLİ
0
-
+
-
+
HSLi
-
+
-
+
-
+
-
+
ATGLİ
-
-
+
+
-
-
+
+
drug (Fig. 5A and 5B). As in H295R cells (Fig. 3C), there was a dose-dependent decrease in staining intensity of triacylglycerol labeled BODIPY (0 uM, 450 vs 50 uM, 268 [P = 0.0682]; 0 µM, 450 vs 100 uM, 150 [P = 0.0015]; 0 µM, 450 vs 200 uM, 69 [P = 0.0003]) (Fig. 5C). Consistent with this, we saw a 2-fold increase in the concentration of gly- cerol in the medium of MUC-1 cells treated with 200 uM mitotane when compared with the control untreated cells or with cells treated with 40 uM mitotane (0 uM, 0.89 mM/µg vs 40 uM, 0.92 mM/ug [P = 0.7142]; 0 uM, 0.89 mM/µg vs 200 uM, 2.4 mM/ug [P < 0.0001]) (Fig. 5D). In contrast to H295R cells, however, we saw a 2.8-fold increase in CE staining in MUC-1 cells treated with 200 uM mitotane, as de- termined by quantification of CE-BODIPY by flow cytometry (0 uM vs 50 uM: 5.6 vs 6.7 [P = 0.9867]; 0 uM vs 100 uM: 5.6 vs 10.11 [P = 0.3964]; 0 uM vs 200 uM: 5.6 vs 16.7 [P = 0.0091]) (Fig. 5E). We confirmed these findings by using imaging flow cytometry, which showed a clear increase in the number and fluorescence intensity of CE-BODIPY-labeled lipid droplets in MUC-1 cells treated with 200 uM mitotane (Fig. 5F). This was accompanied by an increase in SOAT1 and PLIN1 protein levels (Supplementary Figure 2A and 2B (41)).
We reported previously that FC accumulates in MUC-1 cells treated with 200 uM mitotane (33). To determine whether this might be due to induction of LIPE gene expression, we used RT-qPCR to measure LIPE mRNA levels and found a
5-fold increase in MUC-1 cells treated with 200 uM mitotane when compared with the untreated control (0 vs 200 uM, 4.9- fold [P = 0.0198]) (Fig. 5G). There was no effect of mitotane on HSL, PLIN1, or PLIN3 protein levels in MUC-1 cells within normal therapeutic range (Supplementary Figure 2C and 2D). The amount of pHSL was substantially increased by treatment of MUC-1 cells with 200 uM mitotane, resulting in similar amounts to those seen in H295R and HEPG2 cells treated with toxic concentrations of the drug (Fig. 5H). In contrast to H295R cells, however, mitotane induced little or no increase in the amount of total HSL protein in MUC-1 cells (Fig. 5H). The lack of an increase in total HSL protein in MUC-1 cells treated with 200 uM mitotane is surprising given the 5-fold increase in LIPE mRNA but may be explained by changes in protein degradation.
Inhibition of Lipolysis Attenuates Mitotane-induced Cell Death in MUC-1 Cells
Whereas the toxic mechanism of action of mitotane is gener- ally attributed to changes in CEs, the effects we observed on triacylglycerol in mitotane-sensitive H295R cells (Fig. 4) sug- gested an additional, complementary mechanism of toxicity through lipolysis. In mitotane-resistant MUC-1 cells, we also found that supratherapeutic lethal concentrations of mitotane decreased the number of triacylglycerol-labelled lipid droplets and increased levels of the active form of the lipolytic enzyme
A
B
Vehicle Control
Mitotane 50 UM
Mitotane 100 μM
Mitotane 200 μΜ
p=1.51x10-2
1.5-
p=0.4498
Spot Count Score (BODIPY 493/503)
p=0.7441
1.0-
0.5-
0.0-
BODIPY 493/503: Neutral Lipid Droplets
Mitotane (uM)
0
50
100
200
C
MUC-1
D
MUC-1
E
MUC-1
p=3.00×10-4
p=1.50x10-3
p<1.00x104
3
600
p=0.0682
p<1.00x104
Cholestryl Esters (MFI CE-BODIPY)
30-
p=9.10x103
Triacylglycrol (MFI FA-BODIPY)
Glycerol mM/ µg protein
400
2-
20-
p=0.3964
p=0.7142
p=0.9867
200
1-
10-
0
0
40
200
Mitotane (uM)
0
0
Mitotane (uM)
0
50
100
200
0
50
100
200
Mitotane (UM)
F
Vehicle Control
Mitotane 50uM
Mitotane 100µM
Mitotane 200μM
Cholesteryl Ester BODIPY
G
MUC-1
H
8
Relative LIPE mRNA Expression
p=1.98x102
0
20
40
50
0
50
100
200
0
50
100
200
Mitotane (uM)
6
pHSLser563
4
tHSL
2
B Actin
0
VC
Mitotane
H295R
MUC-1
HEPG2
200 μΜ
pHSL. To investigate whether this affects viability in these cells, we treated them with 200 uM mitotane in the presence or absence of HSLi and measured the percentage of dead cells by propidium iodide staining, as before. Mitotane alone
caused a substantial amount of cell death; however, when used in combination with HSLi, as in H295R cells, cell death was significantly attenuated; HSLi alone had no significant effect on MUC-1 viability (VC, 1.9% vs HSL, 2.2% [P > 0.9999];
mitotane, 61% vs mitotane + HSLi, 40.5% [P = 0.0093]) (Fig. 6A). Because HSL protein levels were undetectable at baseline by immunoblot, we confirmed the effect of HSL inhibition in MUC-1 by RT-PCR and demonstrate an 11-fold decrease in LIPE mRNA expression (Supplementary Figure 3A (41),). HSLi alone increased cholesteryl ester fluorescence staining in MUC-1 cells as determined by flow cytometric quantification of CE-BODIPY, and the inclusion of 200 uM mitotane had no appreciable effect on this (HSLi, 889 vs HSLi + mitotane, 1035 [P = 0.2587]) (Fig. 6B). HSLi alone also increased triacylglycerol fluorescence staining in MUC-1 cells and this increase was slightly attenuated when the cells were treated with a combination of HSLi and 200 uM mitotane (HSLi vs HSLi + mitotane: 1199 vs 931 [P = 0.0157]) (Fig. 6C). These data indicate that in MUC-1 cells mitotane predominantly affects triacylglycerol-containing lipid droplets when given at toxic concentrations and that inhibition of HSL partially mitigates this effect.
We investigated further the possible relevance of triacylglycerol depletion for mitotane-induced MUC-1 cell death by using ATGLi, which inhibits lipolysis of triacylglycerol-containing lipid droplets. When used in com- bination with 200 uM mitotane, ATGLi attenuated MUC-1 cell death to a similar extent to HSLi (mitotane, 68% vs ATGLi + mitotane, 48% [P <0.0001]; mitotane, 68% vs HSLi + mitotane, 45% [P < 0.0001]) (Fig. 6D). When
mitotane was combined with both ATGLi and HSLi, no add- itional attenuation of cell death was observed, and ATGLi alone had no effect on MUC-1 cell viability (0 uM, 3% vs ATGLi, 3% [P > 0.9999]) (Fig. 6D). Triacylglycerol levels in lipid-rich HEPG2 cells responded similarly to those in MUC-1 cells when treated with these compounds (Supplementary Figure 3B (41)), but the reduction in mitotane-induced cell death induced by HSLi and ATGLi was even more marked (Supplementary Figure 3C (41), Fig. 6D). This is likely due to the ability of both HSLi and ATGLi to increase triacylglycerol content much more in HEPG2 than they do in H295R or MUC-1 cells. This suggests that mitotane-mediated reduction in lipid droplet number is a novel mechanism of mitotane ac- tion, which can be attenuated using pharmacological inhibi- tors of lipolysis.
Discussion
This study identifies a novel mechanism of mitotane-mediated lipid droplet remodeling. Using 2 validated models of ACC, we demonstrate that mitotane modulates CE and triacylglycerol- containing LDs and that this modulation is associated with cellular toxicity. Specifically, we have demonstrated differing lipid storage phenotypes between mitotane-sensitive H295R and mitotane-resistant MUC-1, whereby the former display predominantly CE storage and the latter a triacylglycerol
A
MUC-1
B
MUC-1
C
MUC-1
80
p=9.30x10-3
p=0.2355
p=1.57x10-2
% Cell Death
1500
1500
p=2.60x10-3
60
Cholestryl Esters (MFI CE-BODIPY)
Triacylglycerol (MFI FA-BODIPY)
1000-
p=0.2587
1000-
40
p=0.9999
500-
500-
20
0
0
0
Mitotane (UM)
0
Mitotane (UM)
0
0
200
200
0
0
200
200
Mitotane (uM)
0
200
200
HSLİ
+
HSLi
-
-
+
-
+
-
+
HSLİ
-
+
-
+
D
MUC-1
p=7.90x103
p=4.27x10-3
80
p=1.12x103
r
Control
p=0.9996
Cell Death (%)
Mitotane 200μΜ
60
40
20
p=0.9999
0
HSLİ ATGLİ
+
-
+
-
+
-
+
+
+
-
-
+
+
storage phenotype. The contrast in lipid storage between both cell lines is associated with distinct LD-associated protein ex- pression levels between both cell lines. Additionally, differing expression of SOAT1, HSL, and DGAT1 consistently favor CE storage in H295R, and triacylglycerol storage in MUC- 1. Overall, mitotane cytotoxicity across each line was asso- ciated with reduced numbers of LDs. In proposing this novel mechanism for mitotane-induced adrenocorticotoxicity, we in turn invite the opportunity to simultaneously target multiple pathways of lipid metabolism in designing future chemotherapeutic regimens for this persistent cancer.
Adrenocortical cells are lipid rich. Previous literature largely focuses on CEs from a high demand as a steroid pre- cursor (9, 25). We newly demonstrate that triacylglycerol LDs are also abundant, particularly in mitotane-resistant MUC-1 cells. In fact, MUC-1 appears to be less dependent on cholesterol utilization, as evidenced by lower expression of both SCARB1 and PLIN2, respectively, associated with HDL uptake and CE storage. Expression of HSL in MUC-1 is strikingly lower than that of H295R. Hormone-sensitive lipase constitutes approximately 97% of all cholesteryl ester hydrolase activity in the adrenal cortex (15). The markedly lower expression of HSL in MUC-1 suggests (1) a lack of de- pendence on this enzyme for cholesterol utilization or (2) that these cells have developed alternative mechanisms to support cholesterol utilization. Previously, we showed that H295R and MUC-1 exhibit similar levels of intracellular free choles- terol (33). In the current study, we have shown lower SOAT1, lower HSL, and lower SCARB1 expression in MUC-1. This indicates alternative or on-demand mechanisms of cholesterol uptake to support steroid synthesis within this steroidogenic cell line, which are not reliant on typical storage/mobiliza- tion pathways. Given the mechanism for mitotane proposed here through liberation of stored CE, this may represent a contributory factor to mitotane resistance and cell survival in MUC-1, and by inference represent a mechanism of mitotane resistance in the clinical scenario. The significance of diverse lipid handling between both mitotane-sensitive and resistant ACC models is consistent with metabolic adaptation, often described in aggressive, drug-resistant, metastatic cancers (42-45). Better understanding of differing lipid metabolism in ACC therefore provides an interesting option for future therapeutics which merits exploration in patients who do not respond to mitotane.
Mitotane is an adrenocorticotoxic pharmacotherapeutic previously shown to inhibit the enzymatic activity of the cholesteryl ester-storage enzyme SOAT1. Specific targeting of SOAT1 using the inhibitor, ATR-101 (Nevanimibe), success- fully induced adrenocorticotoxicity in vitro and in vivo (46). However, singly targeting this pathway was disappointing in a clinical setting (47). We have previously shown that mitotane accumulates FC in H295R and MUC-1 at toxic concentra- tions and now provide additional evidence of a coinciding de- crease in overall LD numbers (33). These findings, supported by downregulation of perilipin expression in H295R, pointed toward a mechanism of cytotoxicity through depleting intra- cellular stores of LD-bound substrate. Marked depletion of LDs in H295R in response to mitotane is accompanied by lower CE and triacylglycerol levels when compared with base- line. Although SOAT1 inhibition may prevent LD formation, it does not fully explain depletion of existing LD, irrespective of the potential rapidity and efficiency of lipid flux. Here, we support a complementary mitotane-induced mechanism
of lipolysis and LD breakdown that supports SOAT1 in- hibition in accumulating FC and promoting adrenocortical lipotoxicity. This is mediated through down-regulation of protective perilipins and an increase in active HSL following mitotane exposure.
It is interesting that previous work has demonstrated significant intracellular accumulation of several free fatty acid species in H295R following mitotane treatment (9). We present data that demonstrate depletion of TAG LDs with concomitant glycerol increase. In this regard, lipolysis of triacylglycerol to toxic free fatty acids may represent an additional cytotoxic mechanism for mitotane at therapeutic doses. Mitotane-resistant MUC1 cells demonstrated an al- tered lipid droplet milieu only at the supratherapeutic, lethal concentrations and LDs remained stable at usually thera- peutic concentrations, where no cell death is observed. It is interesting that mitotane induced triacylglycerol depletion only at lethal doses in MUC1, again suggesting a role for adrenocorticotoxic effects of free fatty acid accumulation in ACC cells. The increase of cholesteryl esters in MUC1 at le- thal mitotane doses is interesting. On 1 hand, this rise was small and should be interpreted in the context of a very low baseline cholesteryl content of these cells. On the other hand, this finding combined with concomitant SOAT1 and PLIN1 upregulation, suggests adaptation of lipid handling of MUC1 cells to alter not only their baseline lipid droplet content, but also their response to mitotane, when compared with their mitotane-sensitive counterparts.
At cytotoxic concentrations, mitotane activated HSL in H295R, MUC-1, and HEPG2 in a dose-dependent manner. The mechanistic relevance of HSL activation was clearly dem- onstrated by the attenuation of mitotane-induced cell death by HSL inhibition for both ACC cell lines. Additionally, inhib- ition of ATGL with associated reduction in TAG breakdown also attenuated mitotane-induced cell death in both ACC cell lines. This demonstrated again the importance of free fatty acid accumulation as a cytotoxic mechanism, particularly in the context of mitotane-resistant MUC1. In fact, the degree to which cell death was equally attenuated by inhibition of HSL and ATGL, without excess cytotoxicity of the combination may in fact speak to even greater importance of free fatty acids rather than FC accumulation, as an adrenocorticotoxic mechanism for mitotane.
Liberation of stored lipid droplet-substrate has not been investigated as a therapeutic strategy in cancer. However, decreased lipolysis is a well-described mechanism of tumor survival (30, 48, 49). We have demonstrated a dynamic inter- play between stored and hydrolyzed lipids within ACC cells, modulated by mitotane at cytotoxic concentrations, which present an interesting approach to isolating novel thera- peutic pathways and which could complement the activity of mitotane or indeed lead to the development of more tolerable alternatives. Overall, these data represent for the first time inhibition of mitotane-induced cell death through a known target in an ACC model.
Lipid metabolism is recognized as a key driver of tumor survival and metastasis across a range of cancers (29, 30, 49-51). Yet, little detail is known about these pathways in ACC. Significant progress made with regard to genetic drivers of disease has highlighted the importance of lipid handling pathways in ACC (52, 53). Recent data from Mohan et al highlights hypermethylation of G0S2 as a predictor of a re- currence and poor prognosis in ACC (54). This is interesting
given that G0S2 is a well characterized negative regu- lator of ATGL-mediated lipolysis (30, 43, 55). Interestingly H295R exhibit hypermethylation of G0S2 whereas MUC-1 are hypomethylated (data not shown), this is functionally in line with low and high TAG levels demonstrated in each cell line, respectively. Given the clinical relevance of G0S2 in stratifying ACC patients, and its mechanistic role as regulator of lipolysis, it is important to further understand lipid droplet breakdown in ACC tumor progression alone and in the con- text of mitotane therapy.
Previous efforts have been made to induce ACC cell death through altering lipid and lipoprotein availability (using HMGCR inhibitors and lipid uptake inhibitors) but only re- cently by targeting other elements of the lipid pathway (9, 10, 35). Lipid handling demonstrates considerable adaptability across the multiple involved pathways (uptake, de novo syn- thesis, storage, liberation, and efflux) that exhibit compensa- tory interplay to provide the necessary substrate required for cellular function and survival (15, 18). We propose that it is necessary to adopt a multihit approach in targeting lipid me- tabolism to overcome this cross-pathway plasticity.
This study investigates the mechanism of lipolysis fol- lowing mitotane treatment in vitro using mitotane-sensitive H295R and mitotane-resistant metastatic MUC-1. These cell lines represent currently validated human cell culture models of ACC (56). Until recently, H295R cells alone represented the standard in vitro model for early mechanistic/therapeutic investigation of ACC. The MUC1 cell line presents an add- itional model of disease, which is mitotane-resistant, meta- static, and steroidogenic (36). As such, this model has, for the first time, facilitated more detailed investigation of mitotane resistance. With regard to newer in vitro models of ACC, it will be interesting to further evaluate lipolysis and lipid ma- chinery in response to mitotane therapy. In particular, those that have been previously exposed to mitotane therapy in ad- vance of tumor resection and subsequent cell line generation, such as TVBF-7, CU-ACC2, and JIL-266 (57-59). Given the significant heterogeneity demonstrated in ACC tumors, investigating cholesterol and lipid metabolism in each of these models may uncover novel therapeutic strategies.
In summary, this study demonstrates a significant difference for lipid handling between mitotane-sensitive and mitotane- resistant ACC cell lines, with a shift to a predominantly triacylglycerol-storage phenotype in mitotane-resistant cells. We also propose a multihit mechanism of mitotane to induce lipotoxicity in ACC cells by targeting lipolysis, in addition to lipid efflux and storage, as previously demonstrated. We have supported this proposal with data demonstrating the effects of mitotane on LD breakdown and the lipolytic pathway. We highlight the importance of lipid handling pathways as thera- peutic targets for ACC. The multihit action of mitotane ex- plains its superior efficacy to agents that singly target aspects of the lipid pathway, such as SOAT1 inhibitors. Finally, these data speak to the need of developing multi-hit therapeutics through single or multiple agents for drug management of ACC, targeting the lipid pathway.
Acknowledgments
The authors thank Ms. Coralie Mureau and Dr. Mark Webber for their technical assistance at The Lambe Institute for Translational Research. The authors thank Shirley Hanley at the Flow Cytometry Core Facility at NUI Galway,
which is funded by NUI Galway, Science Foundation Ireland, the Irish Government’s Programme for Research in Third Level Institutions, Cycle 5, and the European Regional Development Fund. Funding in support of imaging cytometry was received from Science Foundation Ireland under research infrastructure grant no. 16/RI/3760 and from the European Regional Development Fund. We also thank Professor Valentina Boeva at ETH Zurich for helpful computational discussion.
Funding
K.W. is supported by the Irish Research Council Government of Ireland Postgraduate Research Scholarship. E.R.M. is sup- ported by a LifETIME/SFI postgraduate PhD research schol- arship. C.H. is supported by the Uniscientia Foundation (Keyword Tumor Model). This work is also supported by the Irish Endocrine Society Basic Science Research Grant, the Postgraduate Scholarship Fund at the School of Medicine, NUI Galway, and the SFI/NIH US-Ireland Partnership Programme (20/US/3676).
Author Contributions
K.W. and M.C.D. designed the study. K.W., Y.J.L., and E.R. performed the experiments. K.W. and M.C.D. completed data analysis. K.W. and M.C.D. wrote the manuscript. All au- thors have contributed to data interpretation and critical re- vision of the manuscript.
Conflict of Interest
The authors declare that they have no conflict of interest.
Data Availability
The authors confirm that the data supporting the findings of this study are available within the manuscript and the sup- plementary materials. Raw data that support findings of this study are available from the corresponding author upon rea- sonable request.
References
1. Libe R, Borget I, Ronchi CL, et al. Prognostic factors in stage III-IV adrenocortical carcinomas (ACC): an European Network for the Study of Adrenal Tumor (ENSAT) study. Ann Oncol. 2015;26(10):2119-2125. Doi: 10.1093/annonc/mdv329
2. Bilimoria KY, Shen WT, Elaraj D, et al. Adrenocortical carcinoma in the United States: treatment utilization and prognostic factors. Cancer. 2008;113(11):3130-3136. Doi: 10.1002/cncr.23886
3. Margonis GA, Kim Y, Prescott JD, et al. Adrenocortical carcinoma: impact of surgical margin status on long-term outcomes. Ann Surg Oncol. 2016;23(1):134-141. Doi: 10.1245/s10434-015-4803-x
4. Berruti A, Grisanti S, Pulzer A, et al. Long-term outcomes of adjuvant mitotane therapy in patients with radically resected adrenocortical carcinoma. J Clin Endocrinol Metab. 2017;102(4):1358-1365. Doi: 10.1210/jc.2016-2894
5. Terzolo M, Angeli A, Fassnacht M, et al. Adjuvant mitotane treatment for adrenocortical carcinoma. N Engl J Med. 2007;356(23):2372-2380. Doi: 10.1056/NEJMoa063360
6. Terzolo M, Baudin AE, Ardito A, et al. Mitotane levels pre- dict the outcome of patients with adrenocortical carcinoma treated adjuvantly following radical resection. Eur J Endocrinol. 2013;169(3):263-270. Doi: 10.1530/EJE-13-0242
7. Doghman M, Lalli E. Lack of long-lasting effects of mitotane adjuvant therapy in a mouse xenograft model of adrenocortical carcinoma. Mol Cell Endocrinol. 2013;381(1-2):66-69. Doi: 10.1016/j.mce.2013.07.023
8. Daffara F, De Francia S, Reimondo G, et al. Prospective evalu- ation of mitotane toxicity in adrenocortical cancer patients treated adjuvantly. Endocr Relat Cancer. 2008;15(4):1043-1053. Doi: 10.1677/ERC-08-0103
9. Sbiera S, Leich E, Liebisch G, et al. Mitotane inhibits sterol-o-acyl transferase 1 triggering lipid-mediated endoplasmic reticulum stress and apoptosis in adrenocortical carcinoma cells. Endocrinology. 2015;156(11):3895-3908. Doi: 10.1210/en.2015-1367
10. Hescot S, Seck A, Guerin M, et al. Lipoprotein-free mitotane exerts high cytotoxic activity in adrenocortical carcinoma. J Clin Endocrinol Metab. 2015;100(8):2890-2898. Doi: 10.1210/ JC.2015-2080
11. Hescot S, Amazit L, Lhomme M, et al. Identifying mitotane-induced mitochondria-associated membranes dysfunctions: metabolomic and lipidomic approaches. Oncotarget. 2017;8(66):109924- 109940. Doi: 10.18632/oncotarget.18968
12. Cummins CL, Volle DH, Zhang Y, et al. Liver X receptors regulate adrenal cholesterol balance. J Clin Invest. 2006;116(7):1902-1912. Doi: 10.1172/JCI28400
13. Kraemer FB, Shen WJ, Azhar S. SNAREs and cholesterol move- ment for steroidogenesis. Mol Cell Endocrinol. 2017;441:17-21. Doi: 10.1016/j.mce.2016.07.034
14. Kraemer FB. Adrenal cholesterol utilization. Mol Cell Endocrinol. 2007;265-266:42-45. Doi: 10.1016/j.mce.2006.12.001
15. Kraemer FB, Shen WJ, Harada K, et al. Hormone-sensitive lipase is required for high-density lipoprotein cholesteryl ester-supported adrenal steroidogenesis. Mol Endocrinol. 2004;18(3):549-557. Doi: 10.1210/me.2003-0179
16. Hoekstra M, Meurs I, Koenders M, et al. Absence of HDL cholesteryl ester uptake in mice via SR-BI impairs an adequate ad- renal glucocorticoid-mediated stress response to fasting. J Lipid Res. 2008;49(4):738-745. Doi: 10.1194/jlr.M700475-JLR200
17. Goldstein JL, Brown MS. Regulation of the mevalonate pathway. Nature. 1990;343(6257):425-430. Doi: 10.1038/343425a0
18. Kraemer FB, Shen WJ, Patel S, Osuga J, Ishibashi S, Azhar S. The LDL receptor is not necessary for acute adrenal steroidogenesis in mouse adrenocortical cells. Am J Physiol Endocrinol Metab. 2007;292(2):E408-E412. Doi: 10.1152/ajpendo.00428.2006
19. Greenberg AS, Shen WJ, Muliro K, et al. Stimulation of lipolysis and hormone-sensitive lipase via the extracellular signal-regulated kinase pathway. J Biol Chem. 2001;276(48):45456-45461. Doi: 10.1074/jbc.M104436200
20. Kraemer FB, Shen WJ, Natu V, et al. Adrenal neutral cholesteryl ester hydrolase: identification, subcellular distribution, and sex differences. Endocrinology. 2002;143(3):801-806. Doi: 10.1210/ endo.143.3.8693
21. Kimmel AR, Sztalryd C. The perilipins: major cytosolic lipid droplet-associated proteins and their roles in cellular lipid storage, mobilization, and systemic homeostasis. Annu Rev Nutr. 2016;36:471-509. Doi: 10.1146/annurev-nutr-071813-105410
22. Hsieh K, Lee YK, Londos C, Raaka BM, Dalen KT, Kimmel AR. Perilipin family members preferentially sequester to either triacylglycerol-specific or cholesteryl-ester-specific intracellular lipid storage droplets. J Cell Sci. 2012;125(Pt 17):4067-4076. Doi: 10.1242/jcs.104943
23. Lass A, Zimmermann R, Oberer M, Zechner R. Lipolysis - a highly regulated multi-enzyme complex mediates the catabolismof cel- lular fat stores. Prog Lipid Res. 2011;50(1):14-27. Doi: 10.1016/j. plipres.2010.10.004
24. Shen WJ, Patel S, Miyoshi H, Greenberg AS, Kraemer FB. Functional interaction of hormone-sensitive lipase and perilipin in lipolysis. J Lipid Res. 2009;50(11):2306-2313. Doi: 10.1194/jlr. M900176-JLR200
25. Plump AS, Erickson SK, Weng W, Partin JS, Breslow JL, Williams DL. Apolipoprotein A-I is required for cholesteryl ester
accumulation in steroidogenic cells and for normal adrenal steroid production. J Clin Invest. 1996;97(11):2660-2671. Doi: 10.1172/ JCI118716
26. Vahouny GV, Chanderbhan R, Hinds R, Hodges VA, Treadwell CR. ACTH-induced hydrolysis of cholesteryl esters in rat adrenal cells. J Lipid Res. 1978;19(5):570-577.
27. Geng F, Cheng X, Wu X, et al. Inhibition of SOAT1 suppresses glio- blastoma growth via blocking SREBP-1-mediated lipogenesis. Clin Cancer Res. 2016;22(21):5337-5348. Doi: 10.1158/1078-0432. CCR-15-2973
28. Cotte AK, Aires V, Fredon M, et al. Lysophosphatidylcholine acyltransferase 2-mediated lipid droplet production supports colo- rectal cancer chemoresistance. Nat Commun. 2018;9(1):322. Doi: 10.1038/s41467-017-02732-5
29. Li P, Lu M, Shi J, et al. Lung mesenchymal cells elicit lipid storage in neutrophils that fuel breast cancer lung metastasis. Nat Immunol. 2020;21(11):1444-1455. Doi: 10.1038/s41590-020-0783-5
30. Zhang X, Saarinen AM, Hitosugi T, et al. Inhibition of intracellular lipolysis promotes human cancer cell adaptation to hypoxia. Elife. 2017;6:e31132. Doi: 10.7554/eLife.31132
31. Kroiss M, Plonne D, Kendl S, et al. Association of mitotane with chylomicrons and serum lipoproteins: practical implications for treatment of adrenocortical carcinoma. Eur J Endocrinol. 2016;174(3):343-353. Doi: 10.1530/EJE-15-0946
32. Paci A, Hescot S, Seck A, et al. Dyslipidemia causes overestimation of plasma mitotane measurements. Endocrinol Diabetes Metab Case Rep. 2016;2016:150135. Doi: 10.1530/EDM-15-0135
33. Warde KM, Schoenmakers E, Ribes Martinez E, et al. Liver X re- ceptor inhibition potentiates mitotane-induced adrenotoxicity in ACC. Endocr Relat Cancer. 2020;27(6):361-373. Doi: 10.1530/ ERC-20-0031
34. Shawa H, Deniz F, Bazerbashi H, et al. Mitotane-induced hy- perlipidemia: a retrospective cohort study. Int J Endocrinol. 2013;2013:624962. Doi: 10.1155/2013/624962
35. Trotta F, Avena P, Chimento A, et al. Statins reduce intratumor cho- lesterol affecting adrenocortical cancer growth. Mol Cancer Ther. 2020;19(9):1909-1921. Doi: 10.1158/1535-7163.MCT-19-1063
36. Hantel C, Shapiro I, Poli G, et al. Targeting heterogeneity of adrenocortical carcinoma: evaluation and extension of preclin- ical tumor models to improve clinical translation. Oncotarget. 2016;7(48):79292-79304. Doi: 10.18632/oncotarget.12685
37. Qiu B, Simon MC. BODIPY 493/503 staining of neutral lipid droplets for microscopy and quantification by flow cytometry. Bio Protoc. 2016;6(17):e1912. Doi: 10.21769/BioProtoc.1912
38. Lillis AP, Muratoglu SC, Au DT, et al. LDL receptor-related pro- tein-1 (LRP1) regulates cholesterol accumulation in macrophages. PLoS One. 2015;10(6):e0128903. Doi: 10.1371/journal. pone.0128903
39. Connelly MA, Klein SM, Azhar S, Abumrad NA, Williams DL. Comparison of class B scavenger receptors, CD36 and scavenger receptor BI (SR-BI), shows that both receptors mediate high den- sity lipoprotein-cholesteryl ester selective uptake but SR-BI exhibits a unique enhancement of cholesteryl ester uptake. J Biol Chem. 1999;274(1):41-47. Doi: 10.1074/jbc.274.1.41
40. Endemann G, Stanton LW, Madden KS, Bryant CM, White RT, Protter AA. CD36 is a receptor for oxidized low density lipopro- tein. J Biol Chem. 1993;268(16):11811-11816
41. Warde K, Lim Y, Ribes Martinez E, et al. Supplementary File_Warde et al., 2022. Figshare. Deposited July 11, 2022. Figure. https://doi. org/10.6084/m9.figshare.20113982.v1.
42. Zhang Z, Li TE, Chen M, et al. MFN1-dependent alteration of mito- chondrial dynamics drives hepatocellular carcinoma metastasis by glucose metabolic reprogramming. Br J Cancer. 2020;122(2):209- 220. Doi: 10.1038/s41416-019-0658-4
43. Zagani R, El-Assaad W, Gamache I, Teodoro JG. Inhibition of ad- ipose triglyceride lipase (ATGL) by the putative tumor suppressor G0S2 or a small molecule inhibitor attenuates the growth of cancer cells. Oncotarget. 2015;6(29):28282-28295. Doi: 10.18632/ oncotarget.5061
44. Waldschmidt JM, Kloeber JA, Anand P, et al. Single-cell pro- filing reveals metabolic reprogramming as a resistance mech- anism in BRAF-mutated multiple myeloma. Clin Cancer Res. 2021;27(23):6432-6444. Doi: 10.1158/1078-0432.CCR-21-2040.
45. Nimmakayala RK, Leon F, Rachagani S, et al. Metabolic program- ming of distinct cancer stem cells promotes metastasis of pancre- atic ductal adenocarcinoma. Oncogene. 2021;40(1):215-231. Doi: 10.1038/s41388-020-01518-2
46. LaPensee CR, Mann JE, Rainey WE, Crudo V, Hunt SW 3rd, Hammer GD. ATR-101, a selective and potent inhibitor of Acyl- CoA acyltransferase 1, induces apoptosis in H295R adrenocortical cells and in the adrenal cortex of dogs. Endocrinology. 2016;157(5):1775-1788. Doi: 10.1210/en.2015-2052
47. Smith DC, Kroiss M, Kebebew E, et al. A phase 1 study of nevanimibe HCl, a novel adrenal-specific sterol O-acyltransferase 1 (SOAT1) inhibitor, in adrenocortical carcinoma. Invest New Drugs. 2020;38(5):1421-1429. Doi: 10.1007/s10637-020-00899-1
48. Tomin T, Fritz K, Gindlhuber J, et al. Deletion of adipose triglyc- eride lipase links triacylglycerol accumulation to a more-aggressive phenotype in A549 lung carcinoma cells. J Proteome Res. 2018;17(4):1415-1425. Doi: 10.1021/acs.jproteome.7b00782
49. Rozeveld CN, Johnson KM, Zhang L, Razidlo GL. KRAS controls pancreatic cancer cell lipid metabolism and invasive potential through the lipase HSL. Cancer Res. 2020;80(22):4932-4945. Doi: 10.1158/0008-5472.CAN-20-1255
50. Yue S, Li J, Lee SY, et al. Cholesteryl ester accumulation induced by PTEN loss and PI3K/AKT activation underlies human pros- tate cancer aggressiveness. Cell Metab. 2014;19(3):393-406. Doi: 10.1016/j.cmet.2014.01.019
51. Nieman KM, Kenny HA, Penicka CV, et al. Adipocytes pro- mote ovarian cancer metastasis and provide energy for rapid
tumor growth. Nat Med. 2011;17(11):1498-1503. Doi: 10.1038/ nm.2492
52. Assie G, Letouze E, Fassnacht M, et al. Integrated genomic charac- terization of adrenocortical carcinoma. Nat Genet. 2014;46(6):607- 612. Doi: 10.1038/ng.2953
53. Zheng S, Cherniack AD, Dewal N, et al. Comprehensive pan- genomic characterization of adrenocortical carcinoma. Cancer Cell. 2016;30(2):363. Doi: 10.1016/j.ccell.2016.07.013
54. Mohan DR, Lerario AM, Else T, et al. Targeted assessment of G0S2 methylation identifies a rapidly recurrent, routinely fatal molecular subtype of adrenocortical carcinoma. Clin Cancer Res. 2019;25(11):3276-3288. Doi: 10.1158/1078-0432.CCR-18-2693
55. Yang X, Lu X, Lombes M, et al. The G(0)/G(1) switch gene 2 regulates adipose lipolysis through association with adipose tri- glyceride lipase. Cell Metab. 2010;11(3):194-205. Doi: 10.1016/j. cmet.2010.02.003
56. Pinto EM, Kiseljak-Vassiliades K, Hantel C. Contemporary preclin- ical human models of adrenocortical carcinoma. Curr Opin Endocr Metab Res. 2019;8:139-144. Doi: 10.1016/j.coemr.2019.08.009
57. Kiseljak-Vassiliades K, Zhang Y, Bagby SM, et al. Development of new preclinical models to advance adrenocortical carcinoma re- search. Endocr Relat Cancer. 2018;25(4):437-451. Doi: 10.1530/ ERC-17-0447
58. Landwehr LS, Schreiner J, Appenzeller S, et al. A novel patient- derived cell line of adrenocortical carcinoma shows a pathogenic role of germline MUTYH mutation and high tumour mutational burden. Eur J Endocrinol. 2021;184(6):823-835. Doi: 10.1530/EJE-20-1423
59. Rossini E, Tamburello M, Abate A, et al. Cytotoxic effect of progesterone, tamoxifen and their combination in experimental cell models of human adrenocortical cancer. Front Endocrinol (Lausanne). 2021;12:669426. Doi: 10.3389/fendo.2021.669426