SCIENCE sciencemag.orgBy Thomas F. Gajewski and Emily F. HiggsT
umor antigen-specific CD8+ T cells
are a critical component of the anti-
tumor immune response. Many can-
cer patients display evidence of an
endogenous T cell response against
their tumors, yet fail to eliminate
tumors unaided. The failure of spontane-
ous immune-mediated tumor rejection is
thought to be partially due to the action of
negative regulatory mechanisms (immune
checkpoints) that inhibit key functional
properties of tumor-infiltrating T cells ( 1 ).
Checkpoint blockade immunotherapies
have demonstrated notable therapeutic
success by overcoming tumor-induced T
cell inhibition; however, their ef-
ficacy is poor when patients lack
evidence of a spontaneous T cell
response ( 2 , 3 ). Innate immune
agonists may promote priming
and recruitment of tumor-specific
CD8+ T cells and are gaining trac-
tion as a cancer immunotherapy
approach. On page 935 and 993
of this issue, Pan et al. ( 4 ) and Chin et al.
( 5 ), respectively, describe innate immune
agonists that show antitumor activity in
preclinical cancer models.
Antibodies targeting the immune check-
point receptor, programmed cell death
protein 1 (PD-1), or its major ligand, PD-
L1, have been approved by the U.S. Food
and Drug Administration for clinical use
in ~15 different cancer entities ( 6 ). Clinical
benefit has been correlated with the pres-
ence of an activated T cell gene signature
prior to treatment ( 2 ), and following anti–
PD-1 administration, a marked expansion
of tumor-infiltrating CD8+ T cells has
been observed ( 3 ). Despite clinical suc-
cesses, a major subset of cancer patients
lack sufficient T cell inflammation, and
these patients generally do not respond to
checkpoint blockade immunotherapy ( 7 ).
It is thought that triggering productive
T cell–based inflammation within the tu-
mor microenvironment may offer the po-
tential to expand the fraction of patients
benefiting from anti–PD-1 treatment and
other immunotherapies.
One strategy toward this goal has been
to gain an understanding of the fundamen-
tal mechanistic steps involved in spontane-
ous T cell activation and tumor infiltrationwhen it does occur, with the aim of mim-
icking or reproducing those steps in the
cases when it does not occur. In general,
an adaptive immune response (i.e., induc-
tion of a T cell or antibody response) first
requires activation of the innate immune
system, which nonspecifically signals the
presence of “danger” or an outside threat.
Preclinical tumor models revealed that en-
dogenous CD8+ T cell priming (activation)
by innate antigen-presenting cells (APCs)
was markedly reduced in mice deficient for
STING (stimulator of interferon genes) ( 8 ).
Mice lacking STING also showed reduced
cytokine production, including interferon-
b (IFN-b), in response to tumor implan-
tation and failed to reject highly immu-nogenic tumors. These defects were not
observed in mice deficient in other innate
immune pathways, such as specific Toll-
like receptors (TLRs).
The STING pathway is a cytosolic DNA-
sensing pathway, and tumor-derived DNA
could be found within the cytosol of tu-
mor-infiltrating APCs. Cytosolic DNA is
detected within cells when it binds to
cGAS [cyclic guanosine monophosphate
(GMP)–adenosine monophosphate (AMP)
(cGAMP) synthase], which generates
cGAMP, which in turn engages and acti-
vates STING ( 9 ). Signaling downstream
of STING leads to APC activation and in-
flammatory cytokine production, which
subsequently promotes T cell priming and
recruitment ( 10 ). Together, these observa-
tions led to the hypothesis that exogenous
agonists of the STING pathway may have
the potential to trigger de novo innate
immune activation, leading to an adap-
tive immune response that can control tu-
mor growth alone or in combination with
checkpoint blockade immunotherapy.
The first STING agonist investigated
for immunotherapy was the molecule
DMXAA, which had antitumor activity in
preclinical models and was subsequently
determined to interact with the mouseSTING molecule but not human STING
( 11 ). The first generation of human STING
agonists, including MIW815 (ADU-S100)
and MK-1454, have been investigated in
early-phase clinical trials alone and in
combination with anti–PD-1. So far, some
clinical responses to these agonists have
been observed, but only in a minority of
patients (12, 13). Several biological consid-
erations are being explored to understand
mechanisms of response versus resistance.
These include deciphering which immune
cells in the tumor microenvironment must
be present for STING agonists to induce
downstream T cell priming, understanding
the optimal dose and schedule of STING
agonists to avoid overstimulation and
negative regulation, and identi-
fying predictive biomarkers for
clinical activity.
The metabolic instability of
cyclic dinucleotide–based STING
agonists requires them to be ad-
ministered intratumorally. The
constraint for intratumoral ad-
ministration itself has limitations,
because physical issues such as increased
intratumoral pressure, restraints on diffu-
sion of the injected agent, and the impos-
sibility of injecting all metastatic lesions
in an advanced cancer patient are all po-
tential barriers to therapeutic efficacy.
A small number of intravenous STING
agonists have begun evaluation in clini-
cal trials (NCT03843359, NCT04420884,
and NCT04096638), and the next focus
of STING agonist development will likely
be on agonists formulated for systemic
administration, such as those reported by
Pan et al. and Chin et al. (see the figure).
Clinical development of systemically
administered STING agonists needs to
account for several important consider-
ations. One is that systemic administra-
tion may lead to greater toxicity, because
engaging APCs outside the tumor micro-
environment may release high amounts
of IFN-b and other inflammatory cyto-
kines. Chin et al. report that efficacious
doses of SR-717 led to significantly lower
concentrations of serum IFN-b than an-
other recently developed systemic STING
agonist, diABZI-2. Systemic administration
of diABZI-2 also promoted tumor control;
however, diABZI-2 stabilizes STING in its
open conformation, similar to the bacterialCANCERImmunotherapy with a sting
New agonists of an innate immune pathway induce antitumor immunity in mice
“Understanding which innate immune pathway
is functionally relevant... will be paramount
toward optimization of innate immune agonist
combinations with existing immunotherapies.”
21 AUGUST 2020 • VOL 369 ISSUE 6506 921
Published by AAAS