Science - USA (2022-04-22)

INSIGHTS | PERSPECTIVES

GRAPHIC: A. MASTIN/

SCIENCE

science.org SCIENCE

CELL BIOLOGY

Resisting attack by repairing the damage

R emoving membrane pores may help cancer cells survive T cell assault

ByNorma W. Andrews

T

umor cells or cells infected by patho-

gens are thought to resist killing by

cytotoxic T lymphocytes (CTLs), but

the mechanisms involved are incom-

pletely understood. On page 377 of this

issue, Ritter et al. ( 1 ) report that the

membrane-remodeling endosomal sorting

complexes required for transport (ESCRT)

proteins accumulate at cytolytic synapses,

the specialized sites where CTLs bind target

cells and secrete the pore-forming protein

perforin. Using high-resolution imaging and

a clever strategy for selecting cells at early

stages of permeabilization, they show that

ESCRT protein recruitment coincides with

the onset of injury and helps cancer cells re-

sist CTL attack in vitro.

ESCRT proteins drive membrane remodel-

ing and scission events in various processes,

including viral budding and cytokinesis ( 2 ,

3 ). The earlier discovery that they facilitate

wound resealing ( 4 ) suggested a direct path

for lesion removal. ESCRT proteins were de-

tected at plasma membrane sites wounded by

lasers ( 4 , 5 ), but uncertainties remained be-

cause lasers can cause heat-related artifacts

and denature membrane components. Ritter

et al. settle this issue by visualizing ESCRT

proteins precisely at the site of a highly

physiological form of membrane injury. They

also observe membrane protrusions contain-

ing ESCRT proteins within the cytolytic syn-

apse, consistent with a previously proposed

mechanism of membrane repair by vesicle

shedding ( 4 , 5 ). However, the membrane de-

tachment events observed to date could have

resulted from laser-induced damage or cell

death. Consistent with this possibility, Ritter

et al. only directly observed membrane shed-

ding in cells that failed to reseal after CTL

attack. Thus, how plasma membrane lesions

are ultimately removed remains unclear.

Ca2+ influx through cell membrane wounds

triggers the formation of micrometer-scale

blebs as a consequence of cortical cytoskel-

eton rearrangements ( 4 , 6 ). These large

blebs, which display a high density of mem-

brane-associated ESCRT proteins ( 4 ), do not

promote lesion removal because the blebs

retract back into the cell body when mem-

brane integrity is restored ( 4 , 6 ), and cells

reseal normally when blebbing is prevented

( 6 ). It is therefore important, in the context of

CTL–target cell engagement, to distinguish

ESCRT-containing membrane buds that are

responsible for shedding perforin lesions

from reversible membrane perturbations

caused by Ca2+ entry. One strategy might

be to compare the topography of cytolytic

synapses resulting from CTL–target cell en-

gagements of variable potency, as described

by Ritter et al., to identify morphological

features specifically associated with more-

efficient ESCRT-dependent membrane re-

pair. Would target cells engaged by CTLs un-

der conditions that favor ESCRT-dependent

survival show more-homogeneous, perhaps

smaller ESCRT-containing membrane buds

within the cytolytic synapse?

In addition to understanding the mecha-

nism of lesion removal, another challenge

will be to determine whether other path-

ways linked to plasma membrane repair,

such as exocytosis and endocytosis ( 7 ), work

independently of or in concert with ESCRT

recruitment. Although not with the precise

localization shown by Ritter et al. for ESCRT,

exocytosis of lysosomes and rapid endocyto-

sis have been observed in cells attacked by

CTLs ( 8 ), mimicking observations in cells

permeabilized by the bacterial pore-forming

toxin streptolysin O (SLO) ( 6 , 9 ). A subse-

quent study challenged the conclusion that

CTL attack enhances target cell endocytosis

( 10 ), but this discrepancy may be ex-

plained by the different endocytosis

markers used in each study.

Granzyme molecules responsible

for the CTL lethal hit can enter tar-

get cells through perforin pores assembled

at the cell surface ( 10 ). However, it remains

possible that after granzyme entry, perfo-

rin pores are rapidly internalized. Perhaps

this explains why Ritter et al. could not de-

tect pore-like structures on the surface of

target cells by focused ion beam–scanning

electron microscopy (FIB-SEM). It will be

interesting to extend the use of FIB-SEM

to examine the endosomal compartment

of recently permeabilized CTL target cells

to see whether it undergoes the expansion

observed by others ( 6 , 8 , 9 ) and whether it

contains structures that could be attributed

to internalized perforin pores.

In addition to ESCRT proteins, others

have detected exocytosis of lysosomes at

Department of Cell Biology and Molecular Genetics,

University of Maryland, College Park, MD, USA.

Email: [email protected]

Perforin

monomer

Perforin

pore

Granzyme

ESCRT

Lys o s o m a l

enzyme

Lysosome

Cytotoxic T cell

Ta r g e t c e l l

ESCRT recruited

to site of perforin

pores

Ca2+

Ca2+

Ca2+ Ca2+

Budding

of wounded

membrane

Scission of

pore-containing

membrane bud?

Endocytosis

of pores?

Exocytosis

of lysosome

Membrane wounding and repair at the cytolytic synapse

Cytotoxic T cells secrete perforin and granzymes, which trigger target cell death. Target cells resist attack

through endosomal sorting complexes required for transport (ESCRT) proteins, which may promote shedding

of perforin pores. This process may also involve Ca2+-triggered exocytosis of lysosomes and rapid endocytosis.

346 22 APRIL 2022 • VOL 376 ISSUE 6591