Elevated sphingosine-1-phosphate lyase leads to increased metabolism and reduced survival in adrenocortical carcinoma
Jack L. Williams, PhD10 Chris Smith, PhD1 Charlotte Hall, PhD1 Zakaa Khaled, BSc1 Avinaash Maharaj, MBBS, PhD1 Ruth Kwong, MBBS1 James Pittaway, MBBS1 Josefina Casas, PhD2 Laila Parvanta, MBBS3 Tarek Ezzat Abdel-Aziz, MBBS, PhD4 Fausto Palazzo, MBBS5 Teng-Teng Chung, MBBS, PhD4 Leonardo Guasti, PhD1 Lou Metherell, PhD1,D and Rathi Prasad MBBS, PhD1,*+D
1Centre for Endocrinology, Charterhouse Square, Queen Mary University of London, London EC1M 6BQ, United Kingdom
2Research Unit on BioActive Molecules (RUBAM), Department of Biological Chemistry IQAC-CSIC, Barcelona and Liver and Digestive Diseases Networking Biomedical Research Centre (CIBEREHD). ISCIII. Madrid, Spain
3Department of Surgery, St Bartholomew’s Hospital, West Smithfield, London EC1A 7BE, United Kingdom
4Department of Endocrinology, University College London Hospitals NHS Foundation Trust, NW1 2PG London, United Kingdom
5Department of Endocrine and Thyroid Surgery, Hammersmith Hospital, Imperial College London, W12 OHS London, United Kingdom
*Corresponding author: Rathi Prasad, Centre for Endocrinology, Charterhouse Square, Queen Mary University of London, London EC1M 6BQ, United Kingdom. Email: r.prasad@qmul.ac.uk
Abstract
Objective: Adrenocortical carcinomas (ACCs) are invasive tumours arising in the adrenal cortex, and steroidogenic tumours are associated with worse prognostic outcomes. Loss-of-function mutations in sphingosine-1-phosphate lyase (SGPL1) cause primary adrenal insufficiency and as a key degradative enzyme in the sphingolipid pathway, SGPL1 also influences the balance of pro-proliferative and pro-apoptotic sphingolipids. We, therefore, hypothesized increased SGPL1 may be linked to increased disease severity in ACC.
Design: Analyse SGPL1 expression impact on patient survival and adrenal cancer cell phenotype. We analysed two ACC cohorts with survival and corresponding transcriptomic data, focusing on SGPL1 and sphingolipid pathway genes. In vitro, we generated SGPL1-knockout and overexpressing H295R adrenocortical cells to investigate the role of SGPL1 in cell signalling in ACCs.
Results: We found increased expression of several sphingolipid pathway receptors and enzymes, most notably SGPL1 correlated with reduced patient survival in both cohorts. Overexpression of SGPL1 in the H295R cell line increased proliferation and migration while reducing apoptosis, while SGPL1 knockout had the opposite effect. RNA-seq revealed a global increase in the expression of genes in the electron transport chain in overexpressing cells, correlating with increased aerobic respiration and glycolysis. Furthermore, the opposite phenotype was seen in cells lacking SGPL1. We subsequently found the increased proliferation is linked to metabolic substrate availability and increased capacity to use different fuel sources, but particularly glucose, in overexpressing cells.
Conclusions: We, therefore, propose that SGPL1-overexpressing ACC tumours reduce patient survival by increasing fuel usage for anabolism and energy production to facilitate growth and invasion.
Keywords: SGPL1, sphingosine-1-phosphate lyase, sphingolipids, adrenocortical carcinoma
Significance
Adrenocortical carcinomas (ACCs) are associated with overall dismal patient survival, with limited treatment options avail- able. Our studies incorporating analysis of public databases highlight several components of the sphingolipid pathway, most particularly Sphingosine-1-phosphate lyase (SGPL1) as potential therapeutic targets for ACCs. Our in vitro studies show overexpression of SGPL1 alters metabolic capacity and fuel usage in adrenocortical carcinoma cells, driving propensity for proliferation and migration, with SGPL1 knockout conferring a near opposite phenotype. The sphingolipid pathway may present the opportunity to provide bespoke patient treatment in ACC depending on the tumour expression of its key enzymes and receptors.
+ These authors contributed equally to this work.
Background
Adrenocortical carcinoma (ACC) has a global annual inci- dence of 0.7-2 individuals per million.1,2 While rare, overall 5-year survival is poor at approximately 35% and the majority of patients develop associated metastases.3,4 Outcomes re- main dismal with conventional therapies including mitotane, a dicholorodiphenyltrichloroethane derived adrenolytic agent, particularly when associated with distant metastases where 5-year survival can be as low as 13%.3,5 Identifying novel therapeutic targets, therefore, remains critical for this devastating disease.
Sphingolipids are defined by their 18-carbon amino alcohol backbone, modifications of which give rise to a variety of sphingolipid species playing an integral role in membrane biology and as bioactive signalling molecules.6 Sphingolipid intermediates, particularly sphingosine/ceramide and sphingosine-1-phosphate (S1P), have opposing roles in regu- lating cancer cell death and proliferation, respectively.7 The sphingolipid pathway is, importantly, amenable to therapeut- ic manipulation.7 Sphingosine-1-phosphate lyase (SGPL1) carries out the final degradative step in the sphingolipid path- way, irreversible cleavage of S1P, governing the flux of the pathway away from ceramide to S1P (Figure 1A).
SGPL1 is implicated in adrenal disease, with total loss of en- zyme activity and upstream accumulation of sphingolipid in- termediates associated with a multisystemic syndrome incorporating primary adrenal insufficiency (PAI), first de- scribed in 2017.8,9 In vitro modelling of the adrenal insuffi- ciency disease using CRISPR-Cas9 engineered knockout of SGPL1 in the human adrenocortical tumour cell (H295R) line revealed significant reduction in steroidogenesis.10 The disease mechanism has not yet been defined, however accumu- lation in upstream sphingolipids, in particular, sphingosine may directly attenuate the activity of steroidogenic factor 1 (SF-1).11
Sphingolipids play a role in regulating cellular bioenerget- ics12 with a potential impact on cancer cell metabolism, with ceramides, in particular, regulating electron transport chain activity. As SGPL1 can also ‘pull’ the sphingolipid rheostat away from pro-apoptotic ceramide/sphingosine and towards anti-apoptotic S1P, we were prompted to consider the role SGPL1 may play in ACC.
In this study, we interrogate publicly available ACC tran- scriptomic data for the expression of sphingolipid pathway en- zymes and receptors and find a number of associations with survival, most significantly with SGPL1 expression. In determining sphingolipid intermediate flow (Figure 1A), ex- pression of SGPL1 intrinsically influences adrenal tumour cell metabolism thereby impacting tumour progression. Using our original PAI in vitro model, together with a stable SGPL1-overexpressing H295R cell line, we investigate the role of SGPL1 in ACC metabolism, highlighting a novel regu- latory mechanism for ACC tumour progression and a poten- tial therapeutic target.
Methods
Cohort data
For mRNA abundance, 2 independent ACC cohorts were an- alysed; TCGA (The Cancer Genome Atlas) cohort13 which in- cluded 79 ACC cases (https://portal.gdc.cancer.gov/projects/ TCGA-ACC) and the cohort detailed by Assie and
colleagues14 that included 44 ACC cases (Gene Expression Omnibus data set GSE49280). For TCGA cohort, mRNA se- quencing data were extracted from the GDC portal (https:/ portal.gdc.cancer.gov/; accessed October 7, 2020), and all cal- culations were performed on log2 values of FPKM-normalized read counts. Robust Multiarray Average algorithm normalized mRNA abundance was used for the Assie et al. Affymetrix data. The cohorts were divided into 2 equal-size groups ac- cording to the transcript abundance of the genes of interest. Overall survival was defined as time elapsed from the primary resection of ACC to death or last follow-up visit. Event-free survival was further analysed in the TCGA cohort (data avail- able for 73 patients), with patients divided into their original groups as above for each gene of interest. One-way ANOVA testing was undertaken for comparison of COC subtyping with mRNA abundance (log2 values of FPKM normalized read counts). Data relating to hormone production for com- parison with SGPL1 abundance was available for 74 of the 79 patients in the TCGA cohort and all 44 patients in the co- hort detailed by Assìe and colleagues (grouped into hormone vs non-hormone producing). Survival curves were obtained by the Kaplan-Meier method and differences in survival were assessed with the log-rank test.
Crispr-Cas9 knockout of SGPL1
NCI-H295R cells were purchased from ATCC. Steroidogenic potential was monitored by regular cortisol assays and west- ern blots for steroidogenic enzymes. Cells were discarded if they showed a significant drop in steroidogenesis, or at pas- sage 30. H295R cells were cultured according to manufac- turer’s instructions and screened monthly for mycoplasma contamination.
We used the same 4 pairs of oligonucleotides previously used to create SGPL1 knockout HeLa lines,13 as gRNAs for Crispr-Cas9 gene editing in H295R cells. We cloned each pair into pSpCas9(BB)-2A-GFP (PX458), a gift from Feng Zhang (Addgene plasmid #41838) according to the protocol described in.16 Each of the 4 plasmids was then transfected in- dividually into H295R cells using Lipofectamine™M 3000 ac- cording to manufacturer’s instruction (Invitrogen). After 3 days, individual GFP-positive cells were isolated by fluorescence-activated cell sorting and seeded into separate wells in 96-well plates for each plasmid. Colonies were ex- panded and clonal knockout cell lines identified by Sanger se- quencing. Knockout was confirmed by western blotting.
Generation of SGPL1-overexpressing cell lines
pCMV-SGPL1-C-GFPspark® (Origene) was packaged into lentiviral particles in HEK293T cells and concentrated in Vivapsin 20 columns (Sartorius). The concentrate was added to H295R cells and, after confirming successful transduction by fluorescence microscopy, 7 days later, the population of GFP-positive cells was isolated by fluorescence-activated cell sorting and expanded. SGPL1 overexpression was validated by western blotting. For the hygromycin cell line, we cloned a T2A-Hygromycin sequence in place of the C-terminal GFP linker and packaged lentiviral particles as above. Cells were then treated with 100 µg/ml Hygromycin B for 2 weeks, then returned to normal media to recover prior to validation and subsequent experiments.
For further methodology, please see the supplemental data
A
Serine + Palmitoyl-CoA
3-Ketodihydrosphingosine
Sphinganine
Hexadecanal + Phosphoethanolamine
Dihydrosphingomyelin
Dihydroceramide
SGPL1
Sphingomyelin
Ceramide
Sphingosine
SPHK1/2
Sphingosine-1-Phosphate
SGPP1/2
Acylceramide
Ceramide-1-phosphate
Galactosylceramide
Glucosylceramide
B
MKI67
C
NR5A1
Low
**
100
MKI67 expression (log2)
6.
**
100
1
Overall survival (%)
Overall survival (%)
High
n = 79
NR5A1 expression (log2)
10-
4-
Low
2-
:
5
50
High
0
50
n = 79
-2.
0
p <0.0001
-4.
p = 0.0030
0
-6
0
0
1000
2000
3000
4000
5000
COC1
COC2
COC3
0
1000
2000
3000
4000
5000
-5
Time elapsed (days)
Time elapsed (days)
COC1
COC2
COC3
COC subtype
COC subtype
D
SGPL1
Low
E
SGPP1
**
100
High
5-
100
Low
Overall survival (%)
n = 79
SGPL1 expression (log2)
Overall survival (%)
High
SGPP1 expression (log2)
5.
4-
n = 79
4.
3-
3
50
p = 0.0010
2-
50
2.
1-
0-
p = 0.3056
1-
0
-1
0
0
0
1000
2000
3000
4000
5000
Time elapsed (days)
COC1
COC2
COC3
0
1000
2000
3000
4000
5000
COC1
COC2
COC3
Time elapsed (days)
COC subtype
COC subtype
F
SGPP2
G
SPHK1
Low
Low
Overall survival (%)
100
High
SGPP2 expression (log2)
10-
**
6-
Overall survival (%)
100
High
SPHK1 expression (log2)
n = 79
5-
n = 79
4-
2-
50
0
50
0
-5-
p <0.0001
p = 0.0211
-2-
0
-10
0
-4
0
1000
2000
3000
4000
5000
COC1
COC2
COC3
0
1000
2000
3000
4000
5000
COC2
Time elapsed (days)
COC1
COC3
Time elapsed (days)
COC subtype
COC subtype
H
SPHK2
4-
Overall survival (%)
100
Low
- High
SPHK2 expression (log2)
n = 79
2-
50
0-
p = 0.0089
0
-2
.
0
1000
2000
3000
4000
5000
COC1
COC3
Time elapsed (days)
COC subtype
Results
High SGPL1 expression is associated with poor survival prognosis in ACC
We sought to examine how the expression of sphingolipid en- zymes and particularly SGPL1 correlates with survival prob- ability in ACCs. Within the TCGA database, there were 79 patients with ACCs and corresponding RNA-seq expression data. Sorting the patients into high (n=40) and low (n=39) expression for each gene, we found decreased survival associ- ated with high expression of established prognostic/diagnostic biomarkers MKI67 and NR5A1 (SF-1)17-20 (Figure 1B and C). Seventy-six of the 79 patients with RNA-seq data also had cluster of cluster (COC) scoring, incorporating DNA copy number, DNA methylation, RNA expression, and microRNA expression data to classify patients into 1 of 3 clus- ters. Cluster of cluster group I (COCI) have the least aggressive tumours, with only 7% disease progression rate while COCIII are the most aggressive and disease progression rate is 96%.13 MKI67 and NR5A1 expression were both significantly higher in COCIII than COCI (Figure 1B and C). Strikingly, high SGPL1 expression is significantly correlated with lower sur- vival, and expression is increased in the COCIII subgroup (Figure 1D). Sphingosine-1-phoshate phosphatase 1 (SGPP1) expression does not correlate with reduced survival or COC subtype (Figure 1E), while high SGPP2 expression is associated with increased survival and SGPP2 expression is significantly higher in COCI than COCII and COCIII (Figure 1F). In contrast, high SPHK1 and SPHK2 expression correlate with reduced survival, and SPHK2 is expressed sig- nificantly higher in COCIII (Figure 1G and H).
We subsequently analysed SGPL1 expression in a small co- hort of ACC tissues, using nuclear Ki67 staining intensity as a marker of increased proliferation and a proxy for a more ag- gressive tumour (Figure S1A). We found a significant correl- ation of H-scores for SGPL1 and nuclear Ki67 in our patient dataset (Figure S1B). There was no significant difference in SGPL1 expression between cortisol-producing and non- producing tumours (Figure S1C). SGPL1 expression in adre- nocortical adenoma was not analysed in this small series nor has it been described in the literature.
Extracellular S1P signals through a family of 5 g-protein- coupled receptors, sphingosine-1-phosphate receptors 1-5 (S1PR1-5), each coupled to a different subset of G-alpha sub- units with distinct downstream signalling. S1PR1 and S1PR3 are considered pro-tumourigenic, S1PR2 can have both pro- and anti-tumourigenic effects, while the picture is less clear for S1PR4 and S1PR5.21 We found increased S1PR2 and S1PR3 expression correlated with reduced survival in the TCGA ACC cohort (Figure S2B and C). No correlations were found for the other S1P receptors (Figure S2A, D and E).
COCII-III are largely similar to the transcriptomic subtype C1A defined by Assìe and colleagues, 14,22 while C1B is similar to COCI. We stratified our cohort by C1A/C1B subclass, find- ing similar distribution of data as found with COC analyses (Figure S3A-L). The subclass of C1A patients had increased expression of SPGL1, SPHK1, SPHK2, S1PR3, and S1PR5 and reduced expression of SGPP2 and S1PR4.
Event-free survival analysis of TCGA ACC patients reca- pitulated results for overall survival analysis (Figure S4A-L). Given the loss of SGPL1 reduces steroidogenesis, we hypothe- sized increased SGPL1 may be associated with increased ster- oidogenesis.8,9 Steroid hormone-producing ACCs have worse
clinical outcomes than hormone non-producing ACCs, and we found increased SGPL1 expression in steroid-producing tumours (Figure S4M).
We then analysed the dataset from Assie and colleagues, 14 which include 44 ACC patients and corresponding survival and Affymetrix array data. Similar to the TCGA data, we found reduced survival in patients with high SGPL1 and in- creased survival in patients with high SGPP2 (Figure S5). Both datasets seem to indicate enzymes that reduce the pool of ceramide and sphingosine and increase the flow of metabo- lites towards S1P (SGPL1, SPHK1/2) reduce patient survival while those that increase sphingosine and ceramide levels (SGPP2) increase patient survival. We also found reduced sur- vival in patients with higher S1PR2 expression (Figure S5H). S1PR3 was not included in this Affymetrix array data. As with the TCGA cohort, we found increased SGPL1 expression in steroidogenic compared with steroid non-producing ACCs (Figure S5K).
SGPL1 overexpression correlates with increased proliferation and migration in H295R cells
To understand how increased SGPL1 might reduce patient survival, we generated an SGPL1 overexpression cell line us- ing NCI-H295R cells. Wild-type H295R cells were transduced with lentiviral particles facilitating the integration of a gene encoding SGPL1 fused to GFP at the C-terminus, followed by cell sorting for GFP-positive cells. GFP fusion at the C-terminus of SGPL1 does not affect lyase function.23 We also generated a knockout line via Crispr-Cas9 to examine the effect of reduced SGPL1 expression (Figure 2A).
Many pathways affected by S1P signalling through the S1P receptors converge on proliferation (Ras, PI3K). We found increased viability in SGPL1-OE cells compared with wild- type, and significantly reduced viability in SGPL1-KO cells (Figure 2B). Differences in viability may be partly due to changes in the balance of pro- and anti-apoptotic sphingolipids and ceramides. Mass spectrometry for a panel of sphingolipids from cellular lipid fractions revealed increased levels of all sphingolipid intermediates including and upstream of S1P in SGPL1-KO cells (Figure 2C). In contrast, SGPL1-OE cells had reduced ceramide and sphingomyelins compared with wild- type. There was no difference in the level of sphingosine com- pared with wild-type, however, there was a reduction in S1P, confirming the SGPL1-GFP fusion protein is catalytically ac- tive. We found increased phosphoethanolamine (PE) in the SGPL1-KO cells, however, it is worth noting the breakdown of S1P by SGPL1 is not the only method of PE synthesis. There was conversely a lower level of PE in the OE cells, though this may also be due to a rapid turnover of PE to other lipid me- tabolites such as CDP-ethanolamine, phosphatidylethanol- amine or phosphatidylcholine. Studies have shown reduced autophagic flux in response to the loss of SGPL1, partly due to the loss of PE.24 Unlike the previous report in neuronal cells, we found no difference in autophagic flux in SGPL1-KO or OE H295R cells (Figure S6A-J), indicating differences seen in PE (Figure 2C) do not translate to differences in autophagy.
Lowered ceramide levels may reduce apoptotic drive in OE cells. We assessed apoptosis by treating cells with 0, 25, 50, and 100 µg/ml cycloheximide (CHX) for 24 h and performed western blotting for cleaved caspase-3 and cleaved poly-ADP ribose polymerase (PARP). We found a significant reduction in cleaved PARP and cleaved-caspase 3 expressions in the
A
B
KO
WT
OE
2.0
92kDa
SGPL1-C-GFP
KO
Absorbance 595nm
1.5-
**
*
*
WT
*
OE
64kDa
SGPL1
1.0
42kDa
Beta-Actin
0.5-
0.0
24h
48h
72h
96h
C
Sphinganine …
Sphingomyelins
Dihydrosphingomyelin
Sphingosine …
Sphingosine-i-phosphate
Ceramides
…
Phosphoethanolamines
4-
…
4200
…
…
…
.
…
-
s
…
10.
..
4000
..
pol million calls
pmol/milion cells
220
pinolimillion cells
pročmillion pails
158-
prodivision cells
1
pinolimillion caille
1.5
piri/million ctils
1
4000
.-
200-
2000-
100-
To
1200-
sã-
2000
1-
$
=
KO
WT
DE
540
KO
WT
DE
0
KO
WT
DE
4
KO
WT
DE
#
KO
WT
DE
BO
KO
WIT
OE
·
KO
WT
OE
E
PARP
Cleaved PARP
D
KO
WT
OE
2.0-
*
.*
15-
CHX
CHX
Relative Expression
.
CHX
1.5-
Relative Expression
10-
113kDa
1.0-
PARP
0.5
5-
89kDa
I
0.0
0
25
50
100
0
·
25
50
100
KO
~17kDa
cleaved
CHX (pg/ml)
CHX (ug/ml)
WT
caspase-3
Cleaved PARP/Total PARP
Cleaved Caspase 3
OE
.
*
**
Total protein
20-
.
Relative Expression
Relative Expression
4
15
=
10
2
4
5
D
I
0
L
1
·
25
50
100
@
25
50
100
F
CHX (µg/ml)
CHX (pg/ml)
KO
WT
OE
G
Vy
Viable …
E
Necrotic *
02
Early Apoptotic
Late Apoptotic
Q1
02
V610/20-A
.
N
V610/20-A
.
…
..
1
4
0
4
₡
#
*
*
*
L-
1
.
#
2
-
4
“y
3
-
®
.
1
R 670/14-A
1ª
16
NE
1
F 070/14-A
15ª
₱
+
+
₱
+
₫
₱
+
de
®
€
dª
R 670/14-A
H
0h
24h
48h
72h
96h
I
KO
100
**
Relative Closure (%)
KO
80
**
WT
WT
60
OE
40
OE
20
0
24h
48h
72h
96h
OE cells compared at all 3 doses (Figure 2D and E). To assess baseline apoptotis, we performed flow cytometry with annex- in V and propidium iodide (PI). Even in unstimulated cells, there were fewer early and late apoptotic cells in the OE group. These data collectively indicate there is increased viability in
SGPL1 overexpressing cells, which may in part be due to re- duced apoptosis. The reduced viability in the KO cells does not appear to be due to differences in apoptosis.
Among the other pro-tumourigenic pathways downstream of S1PR signalling is migration. Signalling through S1PR3,
in particular, induces alterations in Rac and Rho signalling, both of which may contribute to increased invasiveness.2 21 To assess differences in migration, we seeded cells at full con- fluency and then produced a circular wound via a vacuum pump, as described in25 and monitored wound closure. Cells were pre-treated with 5µg/ml mitomycin C to inhibit prolifer- ation. SGPL1-OE cells filled the wound significantly faster than WT cells, while SGPL1-KO cells migrated significantly slower (Figure 2H and I). Crucially, there were no differences in viability in cells treated with mitomycin C in a parallel ex- periment, indicating the differences are likely due to differen- ces in migratory capacity (Figure 2J).
Transcriptomic analysis highlights the enrichment of calcium handling and oxidative phosphorylation genes in SGPL1-KO cells
Prolonged gain or loss of SGPL1 causes cell- and tissue-type specific changes in the levels of sphingolipids.23,24,26-28 We hy- pothesized this may be due to the transcriptional reprogram- ming of the different cell types in response to SGPL1 levels and consequent lipid imbalance. RNAseq was performed on all three cell lines and principal component analysis revealed the transcriptomic signatures of the biological replicates clus- tered together and in distinct groups (Figure 3A). Using a cut-off of an adjusted P-value less than 0.05, there were 2711 signifi- cantly differently expressed genes between SGPL1-WT and -KO, 2568 between SGPL1-WT and -OE, and 5330 between SGPL1-KO and -OE (Figure 3B and C). Interestingly, we did see some differences in the expression of enzymes in the sphingolipid pathway (Figure S7A). Ceramide synthase 4 and 5 (CERS4/5) were upregulated in OE cells while CERS6 was downregulated. Different ceramide synthase family members are selective for different acyl-CoA substrates and consequently generate different ceramide chain lengths. Differences in expres- sion may reflect differences in substrate availability caused by lipid dysgenesis or alterations in other metabolic pathways.
We analysed significantly different genes by ingenuity path- way analysis (IPA) to determine whether particular pathways were functionally enriched in our comparisons. While there were no obvious candidate pathways that were significantly en- riched in both WT vs. OE and KO vs. OE, calcium signalling was functionally enriched in the KO group in comparison to both OE and WT (Figure 3D), potentially indicating altered cal- cium storage in the KO cells. Examining the significant genes further, we found increased expression of calcium pumps (ATP2A2, ATP2B1) and decreased calcium channels (RYR1, RYR3) in KO cells, all of which regulate the flow of calcium ions between the ER into the cytoplasm (Figure S7B). We as- sessed calcium ER storage by Fura-2-AM assay and stimulating with 1 µM thapsigargin. While there was a significantly smaller calcium release from the ER in the SGPL1-KO compared with the SGPL1-OE cells, neither was different with respect to the wild-type (Figure S7C and D). Equally, there were no differen- ces in the baseline calcium levels (Figure S7E), potentially indi- cating the cytoplasmic calcium levels were similar.
SGPL1 expression drives increased oxidative phosphorylation gene expression and subsequent metabolic rate in H295R SGPL1-OE cells
Gene set enrichment analysis (GSEA) of all genes revealed al- tered expression of genes in the Steroid Hormone Biosynthesis pathway in KO cells compared with WT (Figure 3E),
mimicking what we found in a recent parallel study where we found that SGPL1-KO H295R cells are less steroidogenic. There were no differences in steroidogenic gene expression or cortisol production between WT and OE cells (data not shown), similar to what we showed in Figure S1C.
Instead, we found a striking enrichment in oxidative phos- phorylation genes in the OE group compared with the WT and KO (Figure 3E, red box). Intriguingly, nuclear-encoded ETC genes (e.g. NDUFB10, UQRCQ, and ATP5F1D) were downregulated in the KO but mitochondrial encoded genes (e.g. MT-ND1, MT-ATP6) were upregulated relative to WT, potentially indicating changes in mitochondrial organization, as we have shown in SGPL1-KO Hela cells.27 Western blot using an OXPHOS antibody cocktail (Figure 3F) showed in- creased protein expression of NDUFB8, SDHB, and UQCRC2 (Figure 3G), mirroring the RNA-seq data (Figure S8A-C).
This led us to hypothesize there may be increased respir- ation in the OE cells which could account for the increased proliferative capacity. We used the Seahorse MitoStress Test kit to measure oxidative respiration (Figure 4A). SGPL1-OE cells have higher rates of basal respiration, higher maximal respiration, and ATP production (Figure 4B), perhaps due to the increased expression of the proteins relied on for these processes (Figure 3G). We also find increased basal and max- imal glycolysis in OE cells (Figure 4C and D).
We sought to determine whether increases in metabolic rate and proliferation were due to an increase in fuel usage rate or a modified fuel preference. We grew cells in nutrient- poor media and supplemented with either 17.4 uM D-glucose, 2.5 mM L-glutamine, 100 µM palmitic acid, or a combination of 2 or all 3, and observed the effects on pro- liferation. Growing cells in minimal nutrient media abro- gated all growth differences between WT and OE, while OE cells grew significantly faster in media replete with all 3 fuel sources (Figure 4E and F). KO cells grew marginally but significantly slower even in blank media. Supplementing with either L-glutamine, palmitic acid or both together did not produce differences in growth rate be- tween WT and OE (Figure S9A-C). Only in the presence of D-glucose did we reproduce the differences in growth rate (Figure 4G, Figure S9D and E).
To determine whether OE cells have an increased usage and reliance on D-glucose, we performed substrate oxidation as- says, measuring basal respiration before and after the addition of an inhibitor of glucose, glutamine or free fatty acid metab- olism, and thereafter maximal respiration (Figure 4H-K). We saw significantly higher basal respiration in the OE cells before treatment, as previously (Figure 4L). Treating with BPTES, a glutaminase inhibitor, produced only a modest de- crease in respiration, while Etomoxir, an inhibitor of carnitine palmitoyltransferase-1, reduced respiration significantly more in the KO cells (Figure 4M). UK5099, a mitochondrial pyru- vate carrier inhibitor, produced the greatest decrease in respir- ation in all 3 groups. OE cells were still capable of higher maximal respiration following treatment with BPTES and Etomoxir, but not UK5099 (Figure 4N). These data indicated the difference in proliferation may be due to the increased abil- ity to use glucose to either produce ATP or for anabolism. We performed Seahorse Fuel Flux assays to measure each cell types dependence (how much it relies on that fuel source nor- mally) and capacity (ability to use that fuel source if no others are available) for glucose, fatty acids and glutamine. We found
A
0.6 -
0.4 -
C
Color Key
PC2 (17%)
Genotype
0.2 -
KO
OE
-2
0
2
WT
Row Z-Score
0.0 -
·
-0.2-
-0.25
0.00
PC1 (21.95%)
0.25
0.50
B
WT_v_KO
WT_v_OE
Genes
519
260
347
336
1595
1625
WT-1
WT-2
WT-3
KO-1
KO-2
KO-3
OE-1
OE-2
OE-3
1773
KO_V_OE
D
E
Mineralocorticoid Biosynthesis
Gnrh Signaling Pathway
Insulin Secretion Signaling Pathway
Autoimmune Thyroid Disease
Cholesterol Biosynthesis I
Tyrosine Metabolism
Corticotropin Releasing Hormone Signaling
Type 1 Diabetes Mellitus
Cellular Effects of Sildenafil (Viagra)
Lysine Degradation
Beta Alanine Metabolism
Opioid Signaling Pathway
FDR
Zscore
Steroid Hormone Biosynthesis
Axonal Guidance Signaling
Propanoate Metabolism
0.00
Pulmonary Fibrosis Idiopathic Signaling Pathway
2
Butanoate Metabolism
0.05
0.20
Synaptic Long Term Depression
0
Dorso Ventral Axis Formation
0.50
White Adipose Tissue Browning Pathway
-2
Asthma
RHOGDI Signaling-
Arrhythmogenic Right Ventricular Cardiomyopathy
1.00
Protein Kinase A Signaling
Small Cell Lung Cancer
Pvalue
GNRH Signaling
Phosphatidylinositol Signaling System
ES
7e-08
Prostate Cancer
Sperm Motility
1e-06
0.4
Adherens Junction
Oxytocin Signaling Pathway
1e-04
Endometrial Cancer
0.0
GP6 Signaling Pathway
T
3e-04
Ascorbate and Aldarate Metabolism
-0.4
Dilated Cardiomyopathy Signaling Pathway
Proteasome
-0.8
EIF2 Signaling
Huntingtons Disease
Netrin Signaling
Alzheimers Disease
Hepatic Fibrosis / Hepatic Stellate Cell Activation
Parkinsons Disease
Calcium Signaling
Oxidative Phosphorylation
Ribosome
WTvsKO
WTvsOE
KOVSOE
WTvsKO
WTVSOE
KOVSOE
F
KO
WT
OE
G
- GAPDH
2.5
**
*
*
Relative Expression
KO
- ATP5A
2.0-
WT
· OE
- UQCRC2
1.5
- SDHB
1.0
- COXII
0.5-
-NDUFB8
0.0
NDUFB8
SDHB
UQCRC2
COXI!
ATP5A
WT
KO
OE
A
B
C
D
Mito Stress Test
Glycolysis Stress Test
**
**
OCR (pmol/min/Norm. Unit)
200-
OCR (pmol/min/Norm. Unit)
100
**
*
*
WT
ECAR (mpH/min/Norm. Unit)
KO
60-
ECAR (mpH/min/Norm. Unit)
50-
*
*
**
150
OE
40
40-
100
50
30-
20
20
50
10
0-
0
0
0
20
40
60
80
NMOC
Basal Respiration
Maximal Respiration
Proton Leak
ATP Production
0
20
40
60
80
0
Non-Glycolytic Acidification
Time (minutes)
Time (minutes)
Glycolysis
Glycolytic Capacity
Glycolytic Reserve
+D-Glucose +L-Glutamine
+Palmitic Acid
E
F
G
**
Blank
+ D-Glucose
*
2.0
KO
2.0
2.0
Absorbance 595nm
**
WT
Absorbance 595nm
KO
Absorbance 595nm
KO
1.5
1.5
**
*
WT
OE
1.5.
WT
OE
*
OE
1.0
1.0
*
**
**
**
**
1.0
0.5-
0.5-
0.5
0.0
0.0
0.0
24h
48h
72h
96h
24h
48h
72h
96h
24h
48h
72h
96h
H
DMSO
I
BPTES
J
Etomoxir
K
UK5099
OCR (pmol/min/Norm. Unit)
500
OCR (pmol/min/Norm. Unit)
500-
OCR (pmol/min/Norm. Unit)
500
OCR (pmol/min/Norm. Unit)
500-
KO
400
400
400
400
OE
300
300
300
300
WT
200
200
200
200
100
100-
100-
100-
0
0
0
0
0
30
60
90
120
0
30
60
90
120
0
30
60
90
120
0
30
60
90
120
Time (minutes)
Time (minutes)
Time (minutes)
Time (minutes)
L
M
Acute Response
N
Maximal Respiration
Basal Respiration
*
OCR (pmol/min/Norm. Unit)
250
*
*
AOCR (pmol/min/Norm. Unit)
KO
20
**
KO
OCR (pmol/min/Norm. Unit)
400
*
**
*
KO
200
WT
0
WT
OE
300
WT
150
&
OE
-20.
OE
200-
H
100-
-40
50
-60
100
0
-80
0
DMSO
BPTES
Etomoxir
UK5099
DMSO
BPTES
Etomoxir
UK5099
DMSO
BPTES
Etomoxir
UK5099
O
Glutamine
P
Fatty Acids
Q
Glucose
50
50
100
Fuel Oxidation (%)
Dependency
40
Fuel Oxidation (%)
40
Fuel Oxidation (%)
80
Flexibility
30-
30
60
20
20
40
10-
10.
20-
0
0
0
KO
WT
OE
KO
WT
OE
KO
WT
OE
that the OE cells have a greater capacity (dependency + flexi- bility) for glucose utilization, but not for fatty acids or glutam- ine. This could explain why in the presence of only glutamine
or palmitate, the OE cells grow at the same rate as the WT cells, while in the presence of glucose they are able to grow faster.
Discussion
In recent years pan-genomic characterization has highlighted novel biomarkers for ACC,17,22,29 some of which pose inter- esting therapeutic targets. However, as mitotane is still the only FDA and EMA-approved drug for the treatment of ACC, the search for new therapeutic targets remains critical. A recent study found a network of lipid metabolism genes up- regulated in ACCs, including SGPL1.3º We found a similar re- sult in our studies, while we also found alterations in the expression of other sphingolipid pathway genes (SPHKs/ SGPPs/S1PRs) also correlated with significant differences in survival, as was seen for SPHK1 in a previous study.31 Therefore, the pathway as a whole may represent a possible therapeutic target and crucially, one that has been targeted in cancer and other diseases and for which drugs are already available.31
The sphingolipid pathway is gathering interest in cancer biol- ogy, particularly in view of the purported actions of pro- proliferative S1P.30,32-35 High expression of SGPL1 is uniquely associated with poor survival prognosis in ACC when compared with other cancers in the TCGA data set, indeed in some cancers, it confers an improved prognosis.36-38 Upregulation of SGPL1 expression is however also described in ovarian cancer39 and chemotherapeutic-resistant ovarian tumours.40 SGPL1 plays an integral role in governing the dynamic balance of sphingolipid flux. While high SGPL1 leads to breakdown of S1P, it also in turn promotes catabolismand reduced accumulation of pro-apoptotic ceramides, as is seen in our in vitro ACC model.
SGPL1 may represent a viable therapeutic target in ACC. While total loss of lyase activity is associated with multi- systemic disease, importantly, incomplete pharmacological SGPL1 inhibition has few side effects, as seen in clinical trials for the treatment of rheumatoid arthritis.41 Fingolimod (FTY720), which is currently approved for the management of MS and inhibits S1PR1, 3, & 5, has been shown to have anti- tumour efficacy in vitro and in vivo in several cancer cell mod- els. It is a pro-drug that is phosphorylated primarily by SPHK2 and was recently shown to be efficacious, inhibiting H295R growth in vitro and in vivo.31 Interestingly, it has also been demonstrated as an SGPL1 inhibitor both in vitro and in vivo murine studies. 42 SPHK inhibition, including with isoform se- lective inhibitors, is being targeted in clinical trials for several cancers.43,44 In a clinical study, the use of intravenous Safingol, a putative SPHK inhibitor, in patients with advanced solid tumours stabilized disease in 6 individuals, notably 2 with ACC, one of whom also had a significant reduction in me- tastases.45 Collectively, these studies and ours suggest there is plenty of scope for developing and using drugs targeting various parts of the sphingolipid pathway in patient tumours.
Our data suggest the increased pro-tumourigenic features of SGPL1high ACC tumours may be due, in part, to an increase in cellular respiration rates partly due to an increased capacity to metabolize fuel faster. Cancer cells are known to alter their metabolism, most famously via the Warburg effect, which serves to increase the anabolic capacity of cells in order to in- crease proliferation.46 SGPL1 overexpressing cells had in- creased glycolysis, however, this did not come at the expense of reduced oxidative phosphorylation, rather, it was in- creased. We hypothesize this reflects an increase in the rate of utilization of sugars, for oxidative phosphorylation but po- tentially also for increasing anabolism. An increase in oxida- tive phosphorylation, even in the face of active glycolysis,
has been found in other tumours, including pancreatic ductal adenocarcinoma, leukaemia, lymphomas and endometrial carcinoma, and is an emerging target for cancer therapy.4 Increased mitochondrial DNA content and expression of elec- tron transport chain proteins are amongst the mechanisms de- scribed in tumours with ongoing high OXPHOS activity.47
SGPL1 plays a critical role in Acetyl CoA regulation, which is used for de novo synthesis of ceramides, while the down- stream conversion of ceramide to sphingosine releases free fatty acids48 The increased glycolysis in OE cells may reflect the increased drive for Acetyl CoA generation from glucose to fuel the sphingolipid pathway, as the flux in this pathway is increased by high SGPL1 levels. The sphingolipid pathway can also impact OXPHOS where ceramides can influence the function of the mitochondria, for instance regulating electron transport chain activity and mitochondrial fission.12 In vitro treatment with C16 ceramides, sphingosine or sphinganine inhibits complex IV,49 with C16 ceramides purported to additionally inhibit complex II.50 In these studies, the accumu- lation of ceramides results in a decrease in ATP production. In our OE cells, we found decreases in ceramides, which may al- low for increased complex IV activity and metabolic capacity. It would be of interest to see whether increased cellular meta- bolic activity and sphingolipid profiling of patient-derived ACC mirror our in vitro model, with scope to test the impact of inhibitors of these pathways.
In this study, we find SGPL1 expression is significantly higher in steroidogenic compared with non-steroid producing ACCs (Figure S4M and S5K), where steroidogenic ACC typic- ally confer worse prognosis. 13 While we do not see a difference between steroidogenesis between WT and SGPL1 OE H295R cells, we have previously demonstrated that SGPL1 KO H295R cells have significantly reduced steroidogenic capacity, and the impact of the balance of accumulated sphingolipids on steroidogenesis is yet to be fully determined in this model.10 Sphingolipids, such as ceramide and sphingosine, diminish steroidogenesis with sphingosine directly attenuating steroi- dogenic factor-1 activity.11,51,52 Mitochondrial dysfunction associated with reduced steroidogenesis is also described in SGPL1 deficient patient cells.27 Beyond its impact on cancer cell metabolism, proliferation, and migration, SGPL1 inhib- ition as a therapeutic strategy may also bear the added advan- tage of impeding steroidogenesis in ACC.
In conclusion, we have identified a potential novel regulator of ACCs, whereby increased SGPL1 expression correlates with reduced overall patient survival. This may be due to in- creases in the expression/activity of metabolic enzymes, par- ticularly those required for glucose metabolism. Increased ATP production and anabolisis could facilitate faster cell div- ision which, paired with reduced apoptotic signalling, in- creases proliferation in cells with elevated SGPL1 expression. Future work will aim to validate SGPL1 as a novel prognostic marker in a large independent series of ACCs at protein level and test the ability of inhibitors of the sphingo- lipid pathway to reduce the pro-tumourigenic phenotype of SGPL1high cells and tumours, which we hope will lead to the elucidation of novel insights into the pathogenesis of ACCs and the generation of new therapeutics.
Acknowledgements
We are grateful to the Pathology Facility, the Flow Cytometry Facility, and the Light Microscopy Facility at Charterhouse
Square for their help capturing and analysing data (Core Award C16420/A18066). All acknowledged authors have provided permission to be named in the manuscript.
Supplementary material
Supplementary material is available at European Journal of Endocrinology online.
Funding
This work was supported by Barts Charity grants MGU0361 to L.M. and R.P .; MGU0438 to L.M .; MGU0528 to R.K .; MGU0436 to L.G .; Government of Trinidad and Tobago Research Fellowship to A.M .; Medical Research Council (MRC) UK Clinical Academic Research Partner grant (MR/ T02402X/1) to R.P. and MRC clinical research fellowships (MR/S022155/1) to J.P. and (MR/W015935/1) to R.K.
Conflicts of interests
The authors declare no competing interests.
Ethics approval and consent to participate
ACC specimens were collected from patients undergoing sur- gery at each of St Bartholomew’s, University College and Hammersmith Hospitals, London, after written consent ob- tained from participants and under the study protocol Genetics of endocrine tumours (REC: 06/Q0104/133). This study protocol was approved by Huntingdon Research Ethics Committee. The study was performed in accordance with the Declaration of Helsinki. No individual patient data is presented in this manuscript.
Data availability
All fast q files for this research have been deposited in the NCBI Gene Expression Omnibus (GEO) repository and can be accessed via GSE190177.
References
1. Kebebew E, Reiff E, Duh QY, Clark OH, McMillan A. Extent of disease at presentation and outcome for adrenocortical carcinoma: have we made progress? World J Surg. 2006;30(5):872-878. https:/ doi.org/10.1007/s00268-005-0329-x
2. Kerkhofs TMA, Verhoeven RHA, Van der Zwan JM, et al. Adrenocortical carcinoma: a population-based study on incidence and survival in the Netherlands since 1993. Eur J Cancer Oxf Engl. 2013; 49(11):2579-2586. https://doi.org/10.1016/j.ejca. 2013.02.034
3. Else T, Kim AC, Sabolch A, et al. Adrenocortical carcinoma. Endocr Rev. 2014;35(2):282-326. https://doi.org/10.1210/er. 2013-1029
4. Glenn JA, Else T, Hughes DT, et al. Longitudinal patterns of recur- rence in patients with adrenocortical carcinoma. Surgery. 2019;165(1):186-195. https://doi.org/10.1016/j.surg.2018.04.068
5. Fassnacht M, Johanssen S, Quinkler M, et al. Limited prognostic value of the 2004 International Union Against Cancer staging clas- sification for adrenocortical carcinoma. Cancer. 2009;115(2): 243-250. https://doi.org/10.1002/cncr.24030
6. Gault CR, Obeid LM, Hannun YA. An overview of sphingolipid metabolism: from synthesis to breakdown. Adv Exp Med Biol. 2010;688:1-23. https://doi.org/10.1007/978-1-4419-6741-1_1
7. Ogretmen B. Sphingolipid metabolism in cancer signalling and ther- apy. Nat Rev Cancer. 2018;18(1):33-50. https://doi.org/10.1038/ nrc.2017.96
8. Lovric S, Goncalves S, Gee HY, et al. Mutations in sphingosine-1- phosphate lyase cause nephrosis with ichthyosis and adrenal insuf- ficiency. J Clin Invest. 2017;127(3):912-928. https://doi.org/10. 1172/JCI89626
9. Prasad R, Hadjidemetriou I, Maharaj A, et al. Sphingosine-1-phosphate lyase mutations cause primary adrenal in- sufficiency and steroid-resistant nephrotic syndrome. J Clin Invest. 2017;127(3):942-953. https://doi.org/10.1172/JCI90171
10. Maharaj A, Güran T, Buonocore F, et al. Insights from long-term follow-up of a girl with adrenal insufficiency and sphingosine-1- phosphate lyase deficiency. J Endocr Soc. 2022;6(5):bvac020. https://doi.org/10.1210/jendso/bvac020
11. Urs AN, Dammer E, Kelly S, Wang E, Merrill AH, Sewer MB. Steroidogenic factor-1 is a sphingolipid binding protein. Mol Cell Endocrinol. 2007; 265-266:174-178.
12. Funai K, Summers SA, Rutter J. Reign in the membrane: how com- mon lipids govern mitochondrial function. Curr Opin Cell Biol. 2020;63:162-173. https://doi.org/10.1016/j.ceb.2020.01.006
13. Zheng S, Cherniack AD, Dewal N, et al. Comprehensive pan- genomic characterization of adrenocortical carcinoma. Cancer Cell. 2016;29(5):723-736. https://doi.org/10.1016/j.ccell.2016.04. 002
14. Assié G, Letouzé E, Fassnacht M, et al. Integrated genomic charac- terization of adrenocortical carcinoma. Nat Genet. 2014;46(6): 607-612. https://doi.org/10.1038/ng.2953
15. Gerl MJ, Bittl V, Kirchner S, et al. Sphingosine-1-phosphate lyase deficient cells as a tool to study protein lipid interactions. PLoS ONE. 2016;11(4):e0153009. https://doi.org/10.1371/journal. pone.0153009
16. Ran FA, Hsu PD, Wright J. Genome engineering using the CRISPR-Cas9 system. Nat Protoc. 2013;8(11):2281-2308. https:// doi.org/10.1038/nprot.2013.143
17. Sbiera S, Schmull S, Assie G, et al. High diagnostic and prognostic value of steroidogenic factor-1 expression in adrenal tumors. J Clin Endocrinol Metab. 2010;95(10):E161-E171. https://doi.org/ 10.1210/jc.2010-0653
18. Duregon E, Volante M, Giorcelli J, Terzolo M, Lalli E, Papotti M. Diagnostic and prognostic role of steroidogenic factor 1 in adreno- cortical carcinoma: a validation study focusing on clinical and pathologic correlates. Hum Pathol. 2013;44(5):822-828. https:/ doi.org/10.1016/j.humpath.2012.07.025
19. Libé R. Adrenocortical carcinoma (ACC): diagnosis, prognosis, and treatment. Front Cell Dev Biol. 2015;3:45. https://doi.org/10.3389/ fcell.2015.00045
20. Beuschlein F, Weigel J, Saeger W, et al. Major prognostic role of Ki67 in localized adrenocortical carcinoma after complete resec- tion. J Clin Endocrinol Metab. 2015;100(3):841-849. https://doi. org/10.1210/jc.2014-3182
21. Wang P, Yuan Y, Lin W, Zhong H, Xu K, Qi X. Roles of sphingosine-1-phosphate signaling in cancer. Cancer Cell Int. 2019;19(1):295. https://doi.org/10.1186/s12935-019-1014-8
22. de Reyniès A, Assié G, Rickman DS, et al. Gene expression profiling reveals a new classification of adrenocortical tumors and identifies molecular predictors of malignancy and survival. J Clin Oncol. 2009;27(7):1108-1115. https://doi.org/10.1200/JCO.2008.18. 5678
23. Reiss U, Oskouian B, Zhou J, et al. Sphingosine-phosphate lyase en- hances stress-induced ceramide generation and apoptosis. J Biol Chem. 2004;279(2):1281-1290. https://doi.org/10.1074/jbc. M309646200
24. Mitroi DN, Karunakaran I, Gräler M, et al. SGPL1 (sphingosine phosphate lyase 1) modulates neuronal autophagy via phosphatidy- lethanolamine production. Autophagy. 2017;13(5):885-899. https://doi.org/10.1080/15548627.2017.1291471
25. De Ieso ML, Pei JV. An accurate and cost-effective alternative meth- od for measuring cell migration with the circular wound closure
assay. Biosci Rep. 2018;38(5):BSR20180698. https://doi.org/10. 1042/BSR20180698
26. Colié S, Van Veldhoven PP, Kedjouar B, et al. Disruption of sphingosine 1-phosphate lyase confers resistance to chemotherapy and promotes oncogenesis through Bcl-2/Bcl-xL upregulation. Cancer Res. 2009;69(24):9346-9353. https://doi.org/10.1158/ 0008-5472.CAN-09-2198
27. Maharaj A, Williams JL, Bradshaw T, et al. Sphingosine-1- phosphate lyase (SGPL1) deficiency is associated with mitochon- drial dysfunction. J Steroid Biochem Mol Biol. 2020;202:105730. https://doi.org/10.1016/j.jsbmb.2020.105730
28. Yuan F, Wang Z, Sun Y, et al. Sgpl1 deletion elevates S1P levels, contributing to NPR2 inactivity and p21 expression that block germ cell development. Cell Death Dis. 2021;12(6):1-15. https:// doi.org/10.1038/s41419-021-03848-9
29. 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. https://doi.org/10.1158/1078-0432.CCR-18-2693
30. Subramanian C, Cohen MS. Identification of novel lipid metabolic biomarkers associated with poor adrenocortical carcinoma progno- sis using integrated bioinformatics. Surgery. 2022;171(1):119-129. https://doi.org/10.1016/j.surg.2021.04.049
31. Xu Y, Dong B, Huang J, et al. Sphingosine kinase 1 is overexpressed and promotes adrenocortical carcinoma progression. Oncotarget. 2015;7(3):3233-3244. https://doi.org/10.18632/oncotarget.6564
32. Gupta P, Taiyab A, Hussain A, Alajmi MF, Islam A, Hassan MI. Targeting the sphingosine kinase/sphingosine-1-phosphate signal- ing axis in drug discovery for cancer therapy. Cancers. 2021;13(8):1898. https://doi.org/10.3390/cancers13081898
33. Piazzesi A, Afsar SY, van Echten-Deckert G. Sphingolipid metabol- ism in the development and progression of cancer: one cancer’s help is another’s hindrance. Mol Oncol. 2021;15(12):3256-3279. https://doi.org/10.1002/1878-0261.13063
34. Sattar RSA, Sumi MP, Nimisha A, et al. S1P signaling, its interac- tions and cross-talks with other partners and therapeutic import- ance in colorectal cancer. Cell Signal. 2021;86:110080. https:// doi.org/10.1016/j.cellsig.2021.110080
35. Companioni O, Mir C, Garcia-Mayea Y, LLeonart ME. Targeting sphingolipids for cancer therapy. Front Oncol. 2021;11:4295. https://doi.org/10.3389/fonc.2021.745092
36. Patmanathan SN, Johnson SP, Lai SL, et al. Aberrant expression of the S1P regulating enzymes, SPHK1 and SGPL1, contributes to a migratory phenotype in OSCC mediated through S1PR2. Sci Rep. 2016;6(1):25650. https://doi.org/10.1038/srep25650
37. Oskouian B, Sooriyakumaran P, Borowsky AD, et al. Sphingosine-1-phosphate lyase potentiates apoptosis via p53- and p38-dependent pathways and is down-regulated in colon cancer. Proc Natl Acad Sci USA. 2006;103(46):17384-17389. https://doi. org/10.1073/pnas.0600050103
38. Brizuela L, Ader I, Mazerolles C, Bocquet M, Malavaud B, Cuvillier O. First evidence of sphingosine 1-phosphate lyase protein expres- sion and activity downregulation in human neoplasm: implication for resistance to therapeutics in prostate cancer. Mol Cancer
Ther. 2012;11(9):1841-1851. https://doi.org/10.1158/1535-7163. MCT-12-0227
39. Hibbs K, Skubitz KM, Pambuccian SE, et al. Differential gene ex- pression in ovarian carcinoma. Am J Pathol. 2004;165(2): 397-414. https://doi.org/10.1016/S0002-9440(10)63306-8
40. Bernardini M, Lee CH, Beheshti B, et al. High-resolution mapping of genomic imbalance and identification of gene expression profiles associated with differential chemotherapy response in serous epi- thelial ovarian cancer. Neoplasia N Y N. 2005;7(6):603-613. https://doi.org/10.1593/neo.04760
41. Bagdanoff JT, Donoviel MS, Nouraldeen A, et al. Inhibition of sphingosine 1-phosphate lyase for the treatment of rheumatoid arthritis: discovery of (E)-1-(4-((1R,2S,3R)-1,2,3,4-tetrahydroxy- butyl)-1H-imidazol-2-yl)ethanone oxime (LX2931) and (1R,2S,3R)-1-(2-(isoxazol-3-yl)-1H-imidazol-4-yl)butane-1,2,3,4- tetraol (LX2932). J Med Chem. 2010;53(24):8650-8662.
42. Bandhuvula P, Tam YY, Oskouian B, Saba JD. The immune modu- lator FTY720 inhibits sphingosine-1-phosphate lyase activity. J Biol Chem. 2005;280(40):33697-33700. https://doi.org/10.1074/jbc. C500294200
43. Janneh AH, Ogretmen B. Targeting sphingolipid metabolism as a therapeutic strategy in cancer treatment. Cancers. 2022;14(9): 2183. https://doi.org/10.3390/cancers14092183
44. Pyne NJ, Pyne S. Recent advances in the role of sphingosine 1-phos- phate in cancer. FEBS Lett. 2020;594(22):3583-3601. https://doi. org/10.1002/1873-3468.13933
45. Dickson MA, Carvajal RD, Merrill AH, Gonen M, Cane LM, Schwartz GK. A phase I clinical trial of safingol in combination with cisplatin in advanced solid tumors. Clin Cancer Res. 2011;17(8):2484-2492. https://doi.org/10.1158/1078-0432.CCR- 10-2323
46. Liberti MV, Locasale JW. The Warburg effect: how does it benefit cancer cells? Trends Biochem Sci. 2016;41(3):211-218. https:// doi.org/10.1016/j.tibs.2015.12.001
47. Ashton TM, McKenna WG, Kunz-Schughart LA, Higgins GS. Oxidative phosphorylation as an emerging target in cancer therapy. Clin Cancer Res. 2018;24(11):2482-2490. https://doi.org/10.1158/ 1078-0432.CCR-17-3070
48. Nagarajan SR, Butler LM, Hoy AJ. The diversity and breadth of cancer cell fatty acid metabolism. Cancer Metab. 2021;9(1):2. https://doi.org/10.1186/s40170-020-00237-2
49. Zigdon H, Kogot-Levin A, Park JW, et al. Ablation of ceramide syn- thase 2 causes chronic oxidative stress due to disruption of the mito- chondrial respiratory chain. J Biol Chem. 2013;288(7):4947-4956. https://doi.org/10.1074/jbc.M112.402719
50. Raichur S, Wang ST, Chan PW, et al. CerS2 haploinsufficiency in- hibits ß-oxidation and confers susceptibility to diet-induced steato- hepatitis and insulin resistance. Cell Metab. 2014;20(4):687-695. https://doi.org/10.1016/j.cmet.2014.09.015
51. Lucki NC, Sewer MB. Multiple roles for sphingolipids in steroid hormone biosynthesis. Subcell Biochem. 2008;49:387-412. https://doi.org/10.1007/978-1-4020-8831-5_15
52. Urs AN, Dammer E, Sewer MB. Sphingosine regulates the transcrip- tion of CYP17 by binding to steroidogenic factor-1. Endocrinology. 2006;147(11):5249-5258. https://doi.org/10.1210/en.2006-0355