Science - USA (2021-07-16)

INSIGHTS | PERSPECTIVES

276 16 JULY 2021 • VOL 373 ISSUE 6552 sciencemag.org SCIENCE

GRAPHIC: KELLIE HOLOSKI/

SCIENCE

By Carl Nathan

T

he cytokine interferon-g (IFN-g),

when released by lymphocytes, aug-

ments the capacity of macrophages

and other host cells to kill certain

intracellular protozoa, bacteria, and

viruses (1–3). However, the function

of most of the hundreds of genes induced

by IFN-g is unknown, and biochemical

mechanisms of pathogen inactivation are

incompletely understood. On page 296 of

this issue, Gaudet et al. ( 4 ) identify an ef-

fector mechanism in the human immune

system: production of IFNg–induced apo-

lipoprotein L3 (APOL3) in epithelial cells,

endothelial cells, and fibroblasts that acts

like a detergent, extracting lipids from

membranes of bacteria in the cytosol, kill-

ing the bacteria. Not only does APOL3

dissolve a biophysical boundary—the bac-

terial inner membrane—but the findings

of Gaudet et al. help dissolve conceptual

boundaries about the composition of the

immune system.

In 1882, Metchnikoff launched the field

of immunology with a microscope, a thorn,

and a starfish. Seeing cells gather where the

thorn punctured the larva, he declared that

phagocytic cells defend the host from invad-

ing microbes. Ehrlich favored soluble factors.

By the 1950s, the paradigm was set: Both

cells and soluble factors are required; the im-

mune cells are macrophages, granulocytes,

and lymphocytes; the soluble factors are

chiefly antibodies and proteins collectively

called complement because they are needed

to lyse cells that antibodies tag for destruc-

tion. From the mid-1970s, immunologists dis-

tinguished dendritic cells from macrophages

and denominated lymphocyte subsets, but

the list of cell types in the immune system re-

mained restricted. That the liver is the major

source of complement and other host defense

proteins did not earn hepatocytes member-

ship in the immune system.

The boundary between what is and is

not part of the immune system has recently

sprung more holes with the recognition

that besides the liver, the nervous system

( 5 , 6 ), epithelia (6–8), erythrocytes ( 6 ), and

microbiota (7–9) are key contributors to

mammalian immunity. Gaudet et al. further

integrate diverse cell types into the immune

system by showing that they can deploy an

IFN-g–induced restriction factor against

bacteria in their cytosol (see the figure).

The findings of Gaudet et al. also weaken

a conceptual boundary established in the

early 1990s, when immunologists divided

the immune system into two modes, innate

and adaptive. Innate immunity, it was pro-

posed, kills pathogens directly. Individual

granulocytes and macrophages of the innate

immune system are not clonally distinct, so

large numbers can be mobilized quickly, but

they lack specificity and memory. Adaptive

immunity depends on lymphocyte clones

that proliferate upon encountering a specific

antigen for which they display a receptor.

Persistence of expanded clones with high-

affinity receptors provides memory. This

binary view accommodated some cross-talk:

Dendritic cells present antigen to lympho-

cytes along with signals that prepare them

to proliferate; lymphocytes recognizing an-

tigen secrete IFN-g to instruct macrophages

to increase their killing capacity and make

antibodies that frustrate pathogens’ attacks

or mark them for destruction by phagocytes

or complement. However, recent discoveries

highlight the importance of innate lympho-

cytes that lack antigen receptors. Such cells

can make IFN-g, including at the command

of granulocytes ( 10 ), bypassing adaptive

cells. Lymphocytes that have antigen recep-

tors can kill bacteria ( 11 ) without relying on

innate cells. Persistent epigenetic changes

can endow innate lymphocytes and macro-

phages with memory ( 12 ).

The study of Gaudet et al. adds support

to the following points of view. Every type

of non–disease-causing cell resident in

a human can potentially be a part of the

immune system. Lymphocytes and macro-

phages can work in both innate and adap-

tive modes. Biochemically, cells kill other

cells or themselves in a limited number of

convergently evolved but divergently acti-

vated ways, including by making holes in

membranes (shooting); oxidizing cellular

constituents (burning); and (in)activating

multiple enzymes and channels (poison-

ing). The immune system’s challenge is to

optimally deploy its effectors against faster-

evolving foes, minimize collateral damage,

and help to repair it.

APOL3 alone did not disrupt the inner

membrane of Gram-negative bacteria. Other

IFN-g–induced factors made the bacteria

susceptible to APOL3, including guanylate-

binding protein 1 (GBP1) ( 4 ). This guanosine

triphosphate (GTP)–hydrolyzing enzyme dis-

rupted the outer bacterial membrane, allow-

ing APOL3 access to the inner membrane.

It will be interesting to examine whether

apolipoproteins and GBPs cooperate to kill

mycobacteria, whose maintenance of an

outer lipid layer is analogous to that of Gram-

negative bacteria.

No mutations in APOL3 have been as-

sociated with an immunodeficiency, so it is

unknown if APOL3 makes a nonredundant

contribution to immunity. Loss-of-function

mutations in human genes involved in pro-

duction of, or response to, IFN-g chiefly lead

to increased susceptibility to mycobacterial

infections, but also to increased suscepti-

bility to Salmonella ( 13 ), an organism lysed

IMMUNOLOGY

Rethinking immunology

An interferon-g–induced apolipoprotein lyses bacterial

membranes in the cytoplasm of host cells

Department of Microbiology and Immunology, Weill

Medical College of Cornell University, New York, NY, USA.

Email: [email protected]

Cytokine release

In response to bacterial

infections, activated

lymphocytes secrete IFN-g.

Lysis of bacterium

Once the bacterium is inside the

cell, APOL3 acts as a detergent to

dissolve inner bacterial membranes,

which kills the bacteria.

Production of APOL3

Epithelial cells, endothelial cells, and fibroblasts

induce expression of APOL3 in response to IFN-g.

IFN-g APOL3

nanodiscs

Invading

bacterium

Dissolving bacteria

Upon detection of a bacterial infection, interferon-g

(IFN-g)-induced expression of apolipoprotein L3

(APOL3) in epithelial and endothelial cells, and

fibroblasts, protects these cells from invading bacteria.

0716Perspectives.indd 276 7/9/21 5:20 PM