Nature - USA (2020-09-24)

Nature | Vol 585 | 24 September 2020 | 607

significantly upregulate PUFA-ePEs and PUFA-ePCs (Fig. 4e, Extended
Data Fig. 12j, Supplementary Data 10). Moreover, knockdown of PEX3 or
AGPS reduced the sensitivity of cardiomyocytes to ferroptosis (Fig. 4f,
Extended Data Fig. 12k), suggesting that PUFA-ePL upregulation con-
tributes to heightened ferroptosis sensitivity in cardiomyocytes. In
summary, selective upregulation of PUFA-ePLs is associated with the
ferroptosis-susceptible state in both neuronal and cardiac lineages
(Fig. 4g).
We describe here the importance of peroxisomes and PUFA-ePLs
in modulating susceptibility to ferroptosis in cancer as well as in neu-
ronal and cardiac conditions. Moreover, acquisition of a low-PUFA-ePL

state, promoted by downregulating the expression of ePL biosyn-

thetic enzymes, can lead to ferroptosis evasion and tumour relapse

in ccRCC-bearing hosts. Although plasmalogens—the major PUFA-ePL

type—have been viewed as antioxidants owing to the reactivity of their

alkenyl-ether group^29 , our study describes an additional, pro-ferroptotic

role that is specific to ePLs that have a PUFA at the sn-2 position. Given

that post-mortem brain samples from patients with Alzheimer’s disease

have been found to exhibit markedly reduced plasmalogen levels^30 ,

further investigation into the role of plasmalogen-mediated ferroptosis

in neurodegeneration is warranted. We also note that ferroptosis of

cardiomyocytes has been reported to contribute to chemotherapy- and

a

c

d

–8 –4 048048

0

0.5

1.0

1.5

2.0

2.5

log 2 (day 6/day 0)

C34:1 ePC

C34:3 ePC

C36:4 ePE

C36:3 ePE

C42:11 ePEC38:7 ePE

C34:1 ePC

C34:4 ePC

C34:2 ePC

C34:2 ePE

C36:5 ePE

C38:4 PE

C38:4 PC

C36:4 PE

C22:6 CE

C20:5 CE

C60:12 TAG

C22:5 CE

C40:6 PE

C58:11 TAG

C22:4 CE

C38:5 DAG

–8 –4

0

1

2

3

4

Other lipids

Ether-phospholipids

C34:1 ePC

C34:3 ePCC36:4 ePE

C36:3 ePE

C34:3 ePE

C36:2 ePE

C36:1 ePC

C36:2 ePC

C34:4 ePC

C34:2 ePEC34:1 ePC

PUFA DAG/TAG/CE

C22:6 LPC

C18:3 CE

C56:8 TAG

C22:6 CEC20:5 CE

C22:5 CE

C38:5 DAG

C20:5 LPC

C56:9 TAG

C22:4 CE

C58:9 TAG

C58:10 TAG

e

–1 01

0

0.4

0.8

1.2

Relative viability

SH-SY5Y

Day 6

Day 0

b

0.150

CP CM

(^0) 0.125
20
40
60
80
0.115
Viability (%)
Viability (%)
P = 0.0001
[ML210] (μM)
f g
0
25
50
75
100
125
siNC
siPEX3
siAGPS
0 0.0750.10.115
[ML210] (μM)
0.15
–5.0 –2.502.55.0
0
1
2
3
4
log 2 (CM/CP)
C34:5 ePC
C36:5 ePC-B
C38:6 ePCC38:7 ePC
C40:7 ePC
C36:5 ePE C38:7 ePE
C38:6 ePEC38:5 ePE
C40:7 ePE
C42:11 ePE
C44:13 ePE
PUFA-ether-PL Diacyl-PUFA-PL Other PL
C40:10 PC
C34:3 PC
C38:6 PE
C36:3 PE
Neuronal
progenitors Differentiated neurons
PUFA-ePLs
Ferroptosis sensitivity
CP Mature CM
PUFA-ePLs
Ferroptosis sensitivity
Kidney
epithelial
cells
GPX4-dependent
ccRCC cells
GPX4-independent
ccRCC cells
PUFA-ePLs
AGPAT3
LOOH
Ferroptosis
GPX4
ACSL4
PUFA-ePE/ePC
PUFA
PUFA-CoA
AGPS
FA
FAR1
G3P
AGP LPA
Diacyl-PUFA
-PE/PC
log 2 (day 12/day 0)
Day 0 (parental) Day 12 (differentiated)
SH-SY5Y
β-3-Tubulin DAPI
TMEM189
PUFA-ePLs
–1 01
Day 0
Day 12
Day 12 + Lip-1
SH-SY5Y
P = 0.0001
P = 0.000075
P = 0.0024
P = 0.0165
P = 0.041
P = 0.039
log 10 ([ML210], μM) log 10 ([ML210], μM)
–log
(adj. 10
P)
–log
(adj. 10
P)
Fig. 4 | Neurons and cardiomyocytes acquire increased PUFA-ePLs and gain
sensitivity to ferroptosis during differentiation. a, Fluorescence
microscopy images showing the expression of the neuronal marker β-3-tubulin
in day-0 (parental) and day-12 (differentiated) SH-SY5Y cells. Scale bar, 100 μm.
b, Viability curves of SH-SY5Y cells at the indicated stages after treatment with
ML210 or ML210 + liproxstatin-1 (Lip-1) for 48 h. n = 2 or 4 biologically
independent samples. Data are mean ± s.d. c, Volcano plots showing the
lipidomic profiling of day-0, -6 and -12 differentiating SH-SY5Y cells. n = 3
biologically independent samples for parental and n = 4 for day-6 or day-12
differentiation conditions. d, Bar plots showing relative viability of human
iPS-cell-derived cardiac progenitors (CP) and mature cardiomyocytes (CM)
treated with ML210 for 24 h. n = 6 biologically independent samples. Data are
mean ± s.d., with individual data points shown. e, Volcano plot showing the
phospholipidome of cardiac progenitors (n = 2 biologically independent
samples) and cardiomyocytes (n = 4 biologically independent samples).
f, Bar plots showing the cardiomyocytes transiently transfected with siRNAs
targeting control (siNC), PEX3 or AGPS. n = 3. Data are mean ± s.d., with
individual data points shown. g, Schematic summarizing the peroxisomes and
the PUFA-ePL biosynthesis pathway and their contribution to ferroptosis
susceptibility in cell-state transitions. P values in volcano plots and growth
curves were calculated using two-tailed Student’s t-tests. Multiple-testing
adjustment was performed using the Benjamini–Hochberg method.
a, b, d–f show representative data from experiments performed in duplicate.
LPA, lysophosphatidic acid; G3P, glycerol-3-phosphate; CE, cholesterol esters;
DAG, diacylglycerol; LOOH, phospholipid hydroperoxides.