Abstract
Harnessing the power of immune system to treat cancer has become a core
clinical approach. However, rewiring of intrinsic circuitry by genomic
alterations enables tumor cells to escape immune surveillance, leading
to therapeutic failure. Uncovering the molecular basis of how tumor
mutations induce therapeutic resistance may guide the development of
intervention approaches to advance precision immunotherapy. Here we
report the identification of the Liver Kinase B1 (LKB1)-Inhibitor of
Apoptosis Protein (IAP)- Janus Kinase 1 (JAK1) dynamic complex as a
molecular determinant for immune response of LKB1-mut lung cancer
cells. LKB1 alteration exposes a critical dependency of lung cancer
cells on IAP for their immune resistance. Indeed, pharmacological
inhibition of IAP re-establishes JAK1-regulated Stimulator of
interferon genes (STING) expression and DNA sensing signaling, enhances
cytotoxic immune cell infiltration, and augmentes immune-dependent
anti-tumor activity in an LKB1-mutant immune-competent mouse model.
Thus, IAP-JAK1-targeted strategies, like IAP inhibitors, may offer a
promising therapeutic approach to restore the responsiveness of
immunologically-cold LKB1-mutant tumors to immune checkpoint inhibitors
or STING-directed therapies.
Subject terms: Cancer therapeutic resistance, Intracellular signalling
peptides and proteins, Immune evasion, Cancer immunotherapy
__________________________________________________________________
LKB1-mut has been shown as a major genetic driver of primary resistance
to immune checkpoint inhibitors. Here this group reports the
identification of the LKB1-IAP-JAK dynamic complex as a molecular
determinant for immune response of LKB1-mut lung cancer cells.
Introduction
Impressive clinical activity of immune system-oriented anti-tumor
strategy has led to the rapid rise of immunotherapy as standard cancer
care^[68]1–[69]5. However, the varying primary response rates and
emerging acquired resistance present daunting challenges in expanding
the impact of immunotherapy^[70]6–[71]9. In parallel with the
immune-targeted effort to search for immune checkpoint molecules and
neoantigens, tumor-intrinsic factors have been suggested to modulate
tumor immune-responsiveness^[72]10–[73]12. For example,
loss-of-function mutations in β2-microglobulin and Janus kinases (JAK)
and amplification of Cyclin D1 have been reported in patients resistant
to immunotherapy^[74]13–[75]15. Therefore, understanding how oncogenic
drivers determine the intricate tumor immune response will be critical
for developing biomarkers and modulators to improve immunotherapy
efficacy^[76]16.
Liver kinase B1 (LKB1), also known as Serine/threonine kinase 11, is a
tumor suppressor that regulates AMP-activated protein
kinases^[77]17,[78]18. The LKB1-SIK axis has been shown to suppress the
development of non-small cell lung cancer^[79]19,[80]20. LKB1 mutations
(mut) frequently occur in lung adenocarcinoma (LUAD)^[81]21,[82]22.
LKB1-mut LUAD tumors gain cellular fitness advantage in part through
mTORC1 activation, and subsequent metabolic rewiring and epigenetic
reprogramming^[83]23. However, LKB1-mut LUAD is considered as
undruggable due to the loss-of-function mutations found in tumors and
is resistant to chemotherapy, targeted therapy, and
immunotherapy^[84]9,[85]24–[86]26.
The role of LKB1-mut in shaping the suppressive tumor-immune
microenvironment is emerging^[87]8,[88]9,[89]27–[90]29. It has been
reported that LKB1-mut is a major genetic driver of primary resistance
to immune checkpoint inhibitors^[91]9,[92]30,[93]31. LKB1-mut LUAD has
been characterized to exhibit an immune-suppressive phenotype through
multiple mechanisms. For example, LKB1-mut LUAD cells have altered
expression of proinflammatory cytokines and immune checkpoint
molecules, reprogrammed immune infiltration, and remodeled
extracellular matrix^[94]9,[95]29,[96]32–[97]36. It has been
demonstrated that LKB1-mut LUAD is associated with repressed expression
of stimulator of interferon genes (STING) and corresponding
tumor-intrinsic DNA-sensing innate immune response^[98]27,[99]28. These
LKB1-mut associated immune-cold features have established LKB1 status
as a predictive biomarker for immunotherapy response. However, the
mechanisms underlying the LKB1-mut genotype and immune suppressive
phenotype remain to be established. Therapeutic approaches to target
LKB1-mut LUAD and reverse its immunosuppression are urgently needed.
In this work, to address these challenges, we take a focused
onco-immune interactome mapping approach^[100]37,[101]38 to connect
LKB1 with components of reported cancer immune-response pathways
coupled with a chemical biology approach to examine the
LKB1-mut-created dependency of tumor cells for survival. Both
approaches converge at the LKB1-IAP-JAK interaction complex that
couples the LKB1 status to the innate immunity regulatory function. The
loss of LKB1 function in LKB1-mut cells appears to create the IAP
dependency through a JAK-regulated STING innate immunity pathway.
Modulators of the IAP-JAK regulatory axis, such as the IAP inhibitors,
are shown to restore STING pathway function and enhance the sensitivity
of LKB1-mut cells to immunotherapeutic assault.
Results
Onco-Immune protein-protein interaction (PPI) screening identifies cIAP1 as a
LKB1 binder
LKB1-mut LUAD defines a genetic subset of lung cancer with an
aggressive clinical presentation and therapeutic
resistance^[102]9,[103]24,[104]32. However, the molecular connectivity
bridging the LKB1-mut genotype and immune suppression phenotype remains
elusive. To uncover the molecular mechanisms that determine
LKB1-dictated immune response, we performed a focused oncogenic
immune-regulatory protein-protein interaction (Onco-Immune PPI) mapping
to link LKB1 with immune response regulatory proteins.
We constructed an Onco-Immune gene library that included 85
open-reading-frames encoding proteins with reported immune-regulatory
activities, such as immunogenic cell death, antigen presentation,
innate immunity, and immune checkpoint. Then, we performed an
OncoImmune PPI mapping using a live cell-based BRET^n
(Nanoluciferase-based bioluminescence resonance energy transfer)
technology to discover interactions between Nluc-tagged LKB1 and
Venus-tagged Onco-Immune genes^[105]37. Using a stringent statistical
cutoff of fold-of-change (FOC) ≥ 4.0 and p ≤ 0.001, we prioritized 16
positive Onco-Immune PPI hits that link LKB1 directly with multiple
immune-regulatory pathways, such as the cIAP1-mediated immunogenic cell
death pathway (Fig. [106]1A and Supplementary Data [107]1)^[108]39.
Fig. 1. Onco-Immune PPI profiling reveals the rewired LKB1-cIAP1 axis.
[109]Fig. 1
[110]Open in a new tab
A Scatter plot showing the identification of LKB1-interacting
immune-regulatory protein binding partners. B Immunoblot showing GST-PD
confirmation of LKB1-cIAP1 PPI. Cell lysate from HEK293T cells
co-expressing GST-cIAP1, cIAP2, or XIAP, VF-tagged LKB1 were subjected
to the GST-PD as indicated. C Immunoblot showing endogenous interaction
of LKB1-cIAP1 with the co-IP assay in H1299 cells. D Schematic
illustration of the design of cIAP1 domain truncations. E, F Immunoblot
showing mapping of LKB1-binding domain on cIAP1. Cell lysate from
HEK293T cells co-expressing GST-LKB1 and VF-tagged cIAP1 full-length
(FL), N-terminal truncation (N), C-terminal truncation (C) or BIR
domain truncations (B1, B2 and B3) were subjected to the GST-PD as
indicated. G Immunoblot showing the mapping of cIAP1-binding domain on
LKB1. Cell lysate from HEK293T cells co-expressing GST-cIAP1 and
VF-tagged LKB1 N-terminal truncation (NT,), kinase domain N-lobe
truncation (NL), kinase domain C-lobe truncation (CL), and C-terminal
truncation (CT) were subjected to the GST-PD as indicated. H Immunoblot
showing the LKB1 kinase-dependency of LKB1-cIAP1 PPI. Cell lysate from
HEK293T cells co-expressing GST-cIAP1 and VF-tagged LKB1 WT or K78I
kinase-dead mutant were subjected to the GST-PD as indicated. Before
collecting lysate, cells were serum-starved for overnight. I Schematic
illustration of the LKB1 domain structures and mutations (upper) and a
bar graph showing the PPI signal between cIAP1 and LKB1 WT or
patient-derived mutants from TR-FRET assay (lower). The PPI signal was
expressed as fold-of-change of the TR-FRET signal over the empty vector
control (-) and presented as mean±SD from n = 4 independent
experiments. P-values were calculated by unpaired Student’s t-test with
two-tailed analysis without adjustments comparing with control. J
Representative immunoblot showing PPI signal between cIAP1 and LKB1 WT
or naturally occurring mutant. Cell lysate from HEK293T cells
co-expressing GST-cIAP1 and VF-tagged LKB1 WT or mutant as indicated
were subjected to the GST-pulldown assay. Source data are provided as a
Source Data file. For (B, C, E–H, and J), data are presented as one
representative blot of n = 3 independent experiments.
To validate the LKB1-cIAP1 interaction and determine its structural
basis, we performed orthogonal PPI detection assays and protein domain
truncation studies. First, we confirmed that LKB1 was in complex with
cIAP1 using an affinity pulldown assay (Fig. [111]1B) and a
co-immunoprecipitation assay at the endogenous level (Fig. [112]1C).
Similar interaction was detected between LKB1 and cIAP2, but not XIAP
(Fig. [113]1B). Then, we conducted cIAP1 and cIAP2 domain truncation
studies (Fig. [114]1D) and have localized the PPI interface to the
N-terminal domain of baculoviral IAP repeats (BIR), particularly the
BIR1-2 domain (Fig. [115]1E, F and Supplementary Fig. [116]S1A, B).
Reciprocally, we examined the structural basis of LKB1 for its
interaction with cIAP1. The LKB1-cIAP1 interaction was primarily
mediated by the LKB1 kinase domain, particularly the C-terminal lobe
(the 132-347 fragment) (Fig. [117]1G). Interestingly, the LKB1-cIAP1
PPI was significantly reduced with LKB1 K78I, a catalytically inactive
mutant (Fig. [118]1H). Furthermore, the naturally occurring LKB1
truncation mutants lacking the kinase domain were unable to interact
with cIAP1 (Fig. [119]1I, J). These results suggest that the LKB1-cIAP1
PPI is LKB1 kinase-dependent. Thus cIAP1 may be dissociated from LKB1
in lung cancer cells with LKB1-inactivating mutations and therefore
provide a potential vulnerability for therapeutic targeting.
Chemical screening reveals immune response sensitizers for LKB1-mut cells
Given the largely unknown and complex oncogenic signaling underpinning
LKB1-mut associated immune suppression, we carried out an unbiased
high-throughput immunomodulator phenotypic (HTiP) screen to identify
therapeutic vulnerability created by LKB1 mutations, particularly their
vulnerability to immune-dependent therapeutic assaults.
To examine whether the established HTiP platform^[120]40, featuring in
vitro cancer- and immune-cell co-culture, could recapitulate LKB1-mut
associated immune suppression, we tested the response of LUAD cells
with LKB1-mut or WT, to the immune-cell attack. First, we tested a pair
of isogenic cell lines: parental H1792 with LKB1-WT and the
corresponding LKB1 shRNA-knockdown (KD) cells with a defined and
matched genetic background. Native non-labeled human PBMCs were added
to provide the allogenic immune selection pressure. We found that
parental H1792 cells with LKB1-WT exhibited high sensitivity to the
immune attack with significantly decreased cell viability, whereas the
effect was drastically attenuated in isogenic cells with LKB1-KD
(Fig. [121]2A). Similar results were observed in other pairs of
isogenic LUAD cells, such as H1299 and H1755 cells (Supplementary
Fig. [122]S2A–C) and in a panel of patient-derived cancer cells with
differential LKB1 status (Fig. [123]2B). These results demonstrate that
the HTiP system can recapitulate the LKB1 mutation-associated intrinsic
immune resistance.
Fig. 2. Discovery of IAP inhibitors as immune response sensitizers in
LKB1-mut lung cancer and identification of IAP-STING axis as downstream
effectors of LKB1-cIAP1 PPI.
[124]Fig. 2
[125]Open in a new tab
A Dose-response curve showing cell viability of isogenic LKB1-WT or KD
H1792 cells co-cultured with PBMC. B AUC analysis of PBMC-dose
dependent killing curves of lung cancer cells with LKB1-WT (Calu-1,
H1299, H1792, and H292) or MUT (A549, H1792, H23 and H460). Each dot
represents an individual cell line, and the data were presented as
mean ± SD. C Selectivity of compounds in the immune cell-dependent
killing of H1755 from the primary screening. D IC[50]s of six IAP
inhibitors in H1755 cancer cell alone culture versus co-culture with
PBMC. E Dose-response confirmation of birinapant-induced
immune-dependent killing in additional LKB1-mut LUAD cell lines as
indicated. F Dose-response curves of birinapant-induced CD8^+ T and
CD56^+ NK cell-dependent killing in H1755 cell. G STING gene expression
in H1755 cells cultured alone or co-cultured with immune cells. The
relative mRNA expression was expressed as fold-of-change upon
birinapant (50 nM) treatment over normalized DMSO control. H The
vialibity of H1755 cells cultured alone or co-cultured with PBMC, or
treated with birinapant (100 nM) or in combination with H151 (5 μM). I
The cell viability of isogenic H1755 cells expressing non-targeting
control (ctrl) shRNA or STING-targeting (STING) shRNA cultured alone or
co-cultured with PBMC, or treated with birinapant (100 nM) as
indicated. Immunoblot (lower) showing birinapant-induced STING
expression in control H1755 cells, but not STING knockdown isogenic
cells. J, K Cell viability of stable isogenic H1755 (J) and A549 (K)
cells overexpressing STING cultured alone or co-cultured with PBMC. L
Immunoblot showing indicated proteins of stable isogenic H1755 and A549
cells overexpressing STING by lentiviral transduction using a
pHAGE-STING plasmid. One representative blot from n = 3 independent
experiments. Source data are provided as a Source Data file. For (A, B)
and (D–G), data are presented as mean ± SD from n = 3 independent
experiments; For H-K, data are presented as mean ± SD from n = 4
independent experiments. P-values were calculated by unpaired Student’s
t-test with two-tailed analysis without adjustments.
To identify potential chemical probes that can restore the
responsiveness of LKB1-mut cells to immune selection pressure, we
performed an HTiP screen using a chemogenomic compound library with
well-annotated bioactive compounds^[126]40–[127]43. This screen
revealed three structurally diverse Inhibitors of Apoptosis Protein
(IAP) antagonists, birinapant, BV6 and GDC0152, that enhanced
immune-dependent killing of LKB1-mut cells (Fig. [128]2C, Supplementary
Fig. [129]S2B, and Supplementary Data [130]2). These results reveal IAP
as an intrinsic barrier for immune response. To further examine the
specific impact of targeting IAP for immunomodulation, we tested the
effect of three additional IAP inhibitors (IAPi), AT406, AZD5582, and
LCL161. All six IAPi induced immune cell-dependent selective killing
with potency in the nanomolar range (Fig. [131]2D). Similar immune
dependency on IAP was observed in additional LKB1-mut LUAD cell lines,
such as A549, H157, and H460 (Fig. [132]2E). Importantly, birinapant,
an IAPi, enhanced the response of LKB1-mut LUAD cells to both isolated
CD8^+ T cells and CD56^+ NK cells (Fig. [133]2F). These results reveal
the IAP status as a potential immune-dependent vulnerability in
LKB1-mut LUAD cells, and targeting IAP may enhance the responsiveness
of LKB1-mut tumors to immune-mediated killing. The immune-dependent
anti-tumor activity of IAPi was more profound towards LKB1-KD cells
than that towards the LKB1-WT counterparts (Supplementary
Fig. [134]S2A–C), which were already sensitive to immune killing,
suggesting IAP as a potential tumor-intrinsic factor that controls
immune sensitivity of tumors with mutated LKB1.
IAPi restores tumor intrinsic STING expression in LKB1-mut cells
The results from both the unbiased Onco-Immune PPI mapping
(Fig. [135]1A) and chemical screening (Fig. [136]2C) converge on IAP as
a LKB1 binder and a therapeutic barrier in LKB1-mut LUAD cells. This
physical and functional connectivity between LKB1 and IAP allowed us to
use IAP inhibitors as powerful chemical tools to probe the molecular
mechanisms of the immune suppressive phenotype associated with LKB1
loss. Using IAPi, we examined the biological consequence of the rewired
LKB1-IAP PPI, with a focus on probing the downstream effectors of this
aberrant PPI for its immune response activity.
First, we examined the effect of IAPi on the expression of known
factors involved in LKB1-mut tumor-immune responsiveness, including
STING, IL-1α, IL-6, G-CSF, and PD-L1,
(Fig. [137]2G)^[138]9,[139]28,[140]29,[141]32. Given the potent
immune-dependent anti-cancer activity of IAPi, we reasoned that the
downstream effectors may also oscillate in a similar immune-dependent
manner upon IAPi treatment. From a focused gene expression profiling of
these factors, we found that the STING mRNA level was significantly
increased upon birinapant treatment only in the presence of immune
cells (Fig. [142]2G), whereas the mRNA levels of other factors, such as
IL-1a, IL-6, GM-CSF, and PD-L1, were not significantly changed or in an
immune-cell independent manner (Fig. [143]2G and Supplementary
Fig. [144]S3).
Suppression of tumor intrinsic STING expression was previously
correlated with LKB1 genetic status and the immune resistance
associated with LKB1-mut (Supplementary Fig. [145]S4)^[146]27,[147]28.
To test whether STING contributes to IAPi-enhanced immune
responsiveness, we performed chemical and genetic perturbation studies
to examine the role of STING in birinapant-induced immune-dependent
anti-tumor activity. A pharmacological loss-of-function study showed
that H-151, a specific palmitoylation inhibitor of STING^[148]44, led
to a significant attenuation of birinapant-induced immune-dependent
killing activity in LKB1-mut cells (Fig. [149]2H). Consistently, a
genetic loss-of-function study via shRNA knockdown of tumor
cell-intrinsic STING in LKB1-mut cells significantly blunted
IAPi-induced immune-dependent killing activity (Fig. [150]2I).
Supporting this notion, a genetic gain-of-function study showed that
overexpression of STING in LKB1-mut cells per se did not compromise
cancer cells autonomous viability, but significantly sensitized cancer
cell’s response to PBMC-mediated immune killing (Fig. [151]2J–L).
Together, these results suggest that STING downregulation is a tumor
intrinsic and immune-dependent vulnerability of LKB1-mut lung cancer
cells and IAP inhibitors enhance the immune responsiveness of LKB1-mut
tumors at least in part through the STING expression restoration.
JAK1-STAT1 mediates the IAPi effect in LKB1-mut cells
Given that birinapant-induced STING expression is immune
cell-dependent, we reasoned that there must be immune co-factors
involved in the IAPi effect. To probe the potential immune co-factors,
we performed transcriptomic profiling to compare the differential
expression genes (DEG) from LKB1-mut tumor cells in the absence or
presence of immune cells upon birinapant treatment (Fig. [152]3A). From
the DEG and gene ontology analysis, 1054 immune-dependent genes were
upregulated by birinapant (Supplementary Data [153]3). These DEGs were
enriched in several tumor immune response pathways (Fig. [154]3B). For
example, birinapant activated Type I IFN response signaling
(Fig. [155]3B), a reported STING-downstream event^[156]45,[157]46,
which was suppressed in LKB1-mut LUAD (Supplementary Fig. [158]S5A,
B)^[159]27,[160]28.
Fig. 3. IAP inhibitors synergize with IFNγ to induce STING expression in
LKB1-mut cells.
[161]Fig. 3
[162]Open in a new tab
A DEGs in H1755 cell treated with birinapant (50 nM) in cancer cells
alone culture or co-cultured with PBMC (n = 3 technical replicates). B
Gene ontology analysis showing DEGs-associated top-enriched pathways.
C, D STING mRNA (C) and proteins expression (D) in A549 cells upon
birinapant (500 nM) treatment without (-) or with (+) IFNγ (1 ng/mL)
for 24 h. E, F STING mRNA (E) and protein expression (F) in A549 cells
upon treatment with IAP inhibitors (500 nM) or in combination with IFNγ
(1 ng/mL) for 24 h. G STING expression in A549 STING-HiBiT cells
treated with IAP inhibitors (500 nM) or in combination with IFNγ
(1 ng/mL) for 24 h (n = 4 independent experiments). H Dose-response
curve of birinapant-induced STING expression in A549 STING-HiBiT cells
in the presence of 1 ng/mL IFNγ. I STING expression in isogenic cIAP1
knockdown A549 cells treated with 1 ng/mL IFNγ for 24 h. J, K STING
protein (J) and mRNA (K) expression in A549 cells treated with
birinapant (500 nM) and IFNγ (1 ng/mL) in combination with JAK
inhibitors, oclacitinib (Ocla, 10 μM) and tofacitinib (Tofa, 10 μM), or
RIPK inhibitor, necrostatin-1 (Nec-1, 10 μM). L STING-HiBiT signal in
genetically engineered A549 cells treated with birinapant (500 nM) and
IFNγ (1 ng/mL) in combination with Ocla(10 μM) and Tofa (10 μM) (n = 4
independent experiments). M, N STING expression in isogenic JAK1 (M) or
STAT1 (N) knockdown (KD) A549 cells treated with birinapant (500 nM)
and 1 ng/mL IFNγ for 24 h. O Viability of H1755 cells in cancer cell
alone culture or co-culture with PBMC in combination with Ocla(10 μM)
and Tofa (10 μM). Source data are provided as a Source Data file. For
(C, E, H, K, and O), data are presented as mean ± SD of n = 3
independent experiments. For (D, F, I, J and M, N), data are presented
as one representative blot of n = 3 independent experiments. P-values
were calculated by unpaired Student’s t-test with two-tailed analysis
without adjustments.
In addition, the interferon-gamma (IFNγ) response pathway was activated
by birinapant in a similar immune-dependent manner (Fig. [163]3A, B),
while IFNγ pathway genes were significantly downregulated in LKB1-mut
LUAD samples (Supplementary Fig. [164]S5C). Moreover, the
neutralization of IFNγ, but not TNFα, significantly attenuated
birinapant-induced immune-dependent anti-tumor activity (Supplementary
Fig. [165]S5D). Furthermore, we and others have demonstrated that IAPi
induces IFNγ production from immune cells^[166]40,[167]47. These
results suggest that IFNγ might be an immune co-factor that synergizes
with an IAPi for its immune-dependent STING-induction and anti-tumor
activity.
To test this hypothesis, we investigated the contribution of IFNγ to
birinapant-induced STING restoration. When cancer cells were cultured
alone, birinapant synergized with supplemented IFNγ to induce STING
mRNA and protein expression in LKB1-mut cancer cells (Fig. [168]3C, D
and Supplementary Fig. [169]S6A–D). The induction of STING expression
was likely due to the IAP inhibition effect, as all six IAP inhibitors
exhibited similar synergistic effects with IFNγ on STING (Fig. [170]3E,
F and Supplementary Fig. [171]S6E, F). Such a synergistic effect
between IAPi and IFNγ was further confirmed at the protein level using
a quantitative STING-HiBiT assay^[172]48, which is a genetically
engineered A549 cell-based reporter system for monitoring endogenous
STING expression (Fig. [173]3G and Supplementary Fig. [174]S7). Using
the STING-HiBiT assay, we found that birinapant induced a
dose-dependent increase in STING protein in LKB1-mut A549 cells, with
an EC[50] of ~ 0.2 μM in the presence of IFNγ (Fig. [175]3H). The
on-target effect of IAP inhibitor-induced STING expression was
confirmed by the genetic loss-of-function perturbation of cIAP1
(Fig. [176]3I). These results suggest that IFNγ is an immune co-factor
that synergizes with IAP inhibitors to restore STING expression.
Given that such synergistic effects were also observed at the STING
mRNA level (Fig. [177]3 and Supplementary Fig. [178]S6), it is possible
that a transcriptional regulatory program induced by an IAPi, or in
combination with IFNγ, might be involved. To test this possibility, we
examined the potential involvement of two reported effector pathways
downstream of IFNγ, one pathway mediated by RIPK and the other by JAK,
respectively^[179]49,[180]50. We found that IAPi-induced STING
restoration was significantly reduced upon treatment with JAK
inhibitors, but not RIPK inhibitors, at both protein and mRNA levels
(Fig. [181]3J, K and Supplementary Fig. [182]S6G, H). These results
were confirmed in A549 cells with the engineered STING-HiBiT reporter
(Fig. [183]3L). In support of these results, shRNA knockdown of JAK1
and STAT1 significantly reduced the synergistic effect of IAPi and IFNγ
on restoring STING expression in LKB1-mut cells (Fig. [184]3M, N). JAK
inhibitors abolished the responsiveness of LKB1-mut tumor cells to
IAPi-induced immune-dependent anti-tumor activity (Fig. [185]3O). These
results demonstrate that IAPi-induced STING restoration and
immune-dependent anti-tumor response in LKB1-mut cells occur through
the activation and sensitization of the tumor intrinsic IFNγ-JAK-STAT
pathway.
IAPi reactivates STING-mediated DNA-sensing signaling in LKB1-mut cells
Suppression of STING expression in LKB1-mut LUAD was demonstrated to
lead to immune suppression via impaired cGAS-STING DNA-sensing innate
immune response signaling and subsequent inhibition of the downstream
TBK1-IRF3 pathway, as well as the expression of IRF3 target cytokine
and chemokine genes^[186]27,[187]28. Given that IAP inhibitors restored
STING expression, we next examined whether IAP inhibitors could restore
the STING downstream TBK1-IRF3 signaling and target gene expression.
At the basal level, without supplementation of exogenous dsDNA or
cyclic dinucleotide, birinapant synergized with IFNγ to activate STING
downstream signaling, as evidenced by significantly increased levels of
p-TBK1 and p-IRF3 (Fig. [188]4A) and enhanced expression of target
cytokines and chemokine genes, such as IFNβ, CXCL10 and CCL5
(Fig. [189]4B–D). These results suggest that IAP inhibition not only
restores STING expression but also reactivates STING-mediated
DNA-sensing signaling in response to cytosolic DNA and 2’,3’-cGAMP,
which could be induced by an IAP inhibitor or IFNγ treatment
(Supplementary Fig. [190]S8), or may arise from LKB1-mut tumor
intrinsic genome instability^[191]51.
Fig. 4. IAP inhibitors synergize with IFNγ to induce STING-mediated DNA
sensing pathway activation in LKB1-mut cells.
[192]Fig. 4
[193]Open in a new tab
A Immunoblot showing indicated proteins in A549 cells treated with
birinapant (500 nM) and/or IFNγ (1 ng/mL) for 24 h as indicated. B–D
Bar graphs showing birinapant and IFNγ combination-induced expression
of IFNβ (B), CXCL10 (C), and CCL5 (D) by qPCR in A549 cells treated
with birinapant (500 nM) and/or IFNγ (1 ng/mL) for 24 h as indicated
(n = 3 independent experiments). E Immunoblot showing indicated
proteins in A549 cells treated with poly(dA:dT) (1 μg/mL) for 4 h in
the presence or absence of 24 h pre-treatment with birinapant (500 nM)
and IFNγ (1 ng/mL) combination. F–H Bar graphs showing
poly(dA:dT)-induced expression of IFNβ (F), CXCL10 (G) and CCL5 (H) by
qPCR in A549 cells treated with poly(dA:dT) (1 μg/mL) for 4 h in the
presence or absence of 24 h pre-treatment with birinapant (500 nM) and
IFNγ (1 ng/mL) combination (n = 3 independent experiments). I
Immunoblot showing indicated proteins in A549 cells treated with
poly(dA:dT) (1 μg/mL) for 4 h in the presence of 24 h pre-treatment of
birinapant (500 nM) plus IFNγ (1 ng/mL) in combination with Ocla
(10 μM), Tofa (10 μM), or H151 (5 μM) as indicated. J–L Bar graphs
showing JAK- and STING-dependency of birinapant-induced expression of
IFNβ (J), CXCL10 (K), and CCL5 (L) by qPCR in A549 cells treated with
poly(dA:dT) (1 μg/mL) for 4 h in the presence of 24 h pre-treatment of
birinapant (500 nM) plus IFNγ (1 ng/mL) in combination with Ocla
(10 μM), Tofa (10 μM) or H151 (5 μM) (n = 3 independent experiments).
Source data are provided as a Source Data file. For (A, E, and I), data
are presented as one representative blot of n = 3 independent
experiments. For (B–D, F–H, and J–L), data are presented as mean
values ± SD. P-values were calculated by unpaired Student’s t test with
two-tailed analysis without adjustments.
Upon further STING activation using poly(dA:dT), an exogenous dsDNA, we
found that the combination treatment of birinapant and IFNγ
significantly augmented dsDNA-induced TBK1 and IRF3 phosphorylation
(Fig. [194]4E) and the expression of IFNβ, CXCL10 and CCL5
(Fig. [195]4F–H). The augmentation effect of the TBK1-IRF3 signaling by
the combination of birinapant and IFNγ was not limited to A549 cells.
Similar effects were observed in additional LKB1-mut cells at the basal
level (Supplementary Fig. [196]S9A–D) or with dsDNA stimulation
(Supplementary Fig. [197]S9E–H). In addition, the birinapant and IFNγ
combination treatment also significantly enhanced LKB1-mut cells’
response to ADU-S100, a synthetic cyclic dinucleotide STING agonist
dithio-(RP, RP)-[cyclic[A(2′,5′)pA(3′,5′)p]] (also known as ML RR-S2
CDA, MIW815, or ADU-S100)^[198]52, as shown by increased TBK1 and IRF3
phosphorylation (Supplementary Fig. [199]S9I). Further, we found that a
JAK or STING inhibitor abolished TBK1 and IRF3 phosphorylation
(Fig. [200]3I) and the expression of IFNβ, CXCL10, and CCL5
(Fig. [201]4J–L) induced by dsDNA, indicating the underlying JAK- and
STING-dependency. Altogether, these results suggest that LKB1-mut cells
with STING downregulation have impaired DNA-sensing signaling, which
can be rescued by IAP inhibition through the IFNγ-JAK-STAT
pathway-mediated STING expression.
IAPi induces STING-mediated apoptosis of LKB1-mut cells and chemotaxis of
immune cells in vitro
To explore the functional consequence of IAP inhibitor-induced STING
expression and activation, we examined the fate of LKB1-mut cells and
immune cell infiltration in vitro. STING activation was shown to induce
apoptosis in a context-specific manner^[202]53–[203]56. Given that IAP
inhibitors enhanced STING expression and activation in combination with
IFNγ, we sought to determine whether the combination of an IAP
inhibitor with IFNγ would affect apoptosis of LKB1-mut cells. We found
that birinapant alone did not lead to significant cell death, while the
combination of birinapant with IFNγ induced significant caspase3/7
activation in LKB1-mut A549 cells (Fig. [204]5A, B). Similar effects of
the birinapant and IFNγ combination on caspase3/7 activation were also
observed in additional LKB1-mut cells (Supplementary Fig. [205]S10A,
B). Such a combination effect was significantly attenuated upon STING
inhibition by H151 (Fig. [206]5C, D). These results reveal that IAP
inhibition can synergize with IFNγ to induce apoptosis of LKB1-mut
tumors in a STING-dependent manner.
Fig. 5. Birinapant induces STING-mediated apoptosis of LKB1-mut cancer cells
and chemotaxis of immune cells in vitro.
[207]Fig. 5
[208]Open in a new tab
A Time-dependent curve of apoptotic cell counts of A549 cells treated
with birinapant (500 nM), IFNγ (1 ng/mL) or in combination. (n = 3
independent biological replicates using different passages of A549
cells). B Representative images showing apoptotic A549 cells treated
with birinapant (500 nM), IFNγ (1 ng/mL), or in combination. Scale bar:
100 μm. C Bar graph showing birinapant-induced STING-mediated cancer
cell apoptosis. A549 cells were treated with birinapant (500 nM), IFNγ
(1 ng/mL), H151 (5 μM,) or in combination as indicated for 72 h. (n = 4
independent experiments). D Representative images showing apoptotic
A549 cells treated with birinapant (500 nM), IFNγ (1 ng/mL), H151
(5 μM), or in combination as indicated for 72 h. Scale bar: 100 μm. E
Schematic illustration of transwell assays for measuring immune cell
infiltration in vitro. F Representative images showing
birinapant-induced Jurkat T cells migration. A549 cells (red) and
Jurkat T cells (green) were co-cultured in transwell as shown in (E)
and were treated with birinapant (100 nM), poly(dA:dT) (1 μg/mL), or in
combination with H151 (5 μM), or JAK inhibitor (JAKi), tofacitinib
(10 μM), for 48 h, labeled as the letters a-h. A549 was labeled with
Nuclight Red fluorescence protein, and Jurkats were pre-labeled with
CellTracker™ Green CMFDA Dye. IL2 and anti-CD3 antibody were used to
activate Jurkat T cells. Scale bar: 200 μm. G Bar graph showing the
quantification of infiltrated Jurkat T cells in transwell-based
migration assays. The letters a-h in lowcase were corresponding to the
conditions of F (n = 3 independent experiments). Source data are
provided as a Source Data file. For (A, C, and G), data are presented
as mean ± SD. P-values were calculated by unpaired Student’s t test
with two-tailed analysis without adjustments.
Given that STING-mediated CXCL10 and CCL5 chemokines are involved in
immune cell chemotaxis^[209]57,[210]58, we examined the potential
functional consequence of IAP inhibitor-induced STING expression on
immune cell chemotaxis in a 2D transwell assay. As shown in
Fig. [211]5E, the birinapant treatment significantly promoted the
migration of IL2- and anti-CD3-activated Jurkat T cells from the upper
chamber to the cancer cell culture in the bottom chamber (Fig. [212]5F,
G and Supplementary Fig. [213]S10C). Similar effects of
birinapant-induced immune cell migration were observed for CD56^+
NK92-MI cells, or with additional dsDNA stimulation (Fig. [214]5F, G
and Supplementary Fig. [215]S10D–F). However, a JAK inhibitor or STING
antagonist abolished the birinapant-induced immune cell migration
(Fig. [216]5F, G and Supplementary Fig [217]S10D–F). These results
suggest that IAP inhibitor-induced STING expression and reactivation
can augment immune cell migration in vitro.
IAP inhibitor induced tumor immune response in vivo in LKB1-mut mouse models
To determine the functional consequence of targeting the rewired
LKB1-cIAP1-JAK1 trimolecular complex in an in vivo setting, we examined
the anti-tumor activity of IAP inhibitors using our developed LKB1-mut
mouse model^[218]36,[219]59. First, we found that WRJ388, a mouse
tumor-derived cancer cell line isolated from our
Kras^G12D/Lkb1^-/-/p53^WT GEMM^[220]36, has downregulated STING
expression as compared to KW634 cells^[221]60
(Kras^G12D/Lkb1^WT/p53^-/-) (Supplementary Fig. [222]S11A, B).
Moreover, IAP inhibitors in combination with mouse IFNγ significantly
induced STING expression (Supplementary Fig. [223]S11C, D) and CCL5
expression (Supplementary Fig. [224]S11E) in Lkb1-mut WRJ388 cells.
These results suggest that WRJ388 cells can recapitulate not only the
LKB1-mut and STING-downregulation genotype-phenotype relationship
observed in human LUAD patients but also the IAP dependency as shown in
human LKB1-mut cells.
Using the allograft model with WRJ388 cells, we assessed the antitumor
activity of birinapant in LKB1-mut background in vivo. Murine WRJ388
cells were subcutaneously injected into the immunocompetent mice,
followed by birinapant treatment (Fig. [225]6A). Treatment with
birinapant led to a ~ 42% reduction in tumor volume from ~ 300 to
~ 175 mm^3, which is ~ 4.4-fold less than the vehicle control at the
endpoint, indicating a significant anti-tumor effect (Fig. [226]6B). By
contrast, no significant tumor shrinkage was observed with the
birinapant treatment in the immune-deficient nude mice (Fig. [227]6C).
Therefore, birinapant exhibits immune-dependent anti-tumor activity in
vivo.
Fig. 6. Birinapant exhibits immune-dependent anti-tumor activity in vivo in a
LKB1-mut syngeneic allograft mouse model.
[228]Fig. 6
[229]Open in a new tab
A Schematics of mouse study design. B, C Time course of tumor volume
change in immune-competent mice (B, n = 6 mice per group) and
immune-deficient nude mice (C, n = 5 mice per group) treated with
vehicle control or birinapant (10 mg/kg) as indicated. The data are
presented as mean ± SD of the entire experimental cohort. D
Representative images from immunohistochemistry (IHC) staining showing
2-doses birinapant-induced increase of tumor infiltrated CD8^+ T cells
in vivo. Tumor samples were collected on Day 6 for the IHC staining
analysis. Scale bar: 100 μm. E tSNE plots from single-cell mass
cytometry (CyTOF) showing increased tumor infiltrated CD8^+ T cells
upon 2-doses birinapant treatment in vivo. (F) Quantification of
tumor-infiltrated CD8^+ T cells from single-cell CyTOF profiling. The
data are expressed as the percentage of CD8^+ T cells in the live cell
population. Each data point represents individual samples from
immune-competent mice. (n = 5 mice for vehicle group, n = 7 mice for
birinapant group). G Immunoblot showing birinapant-induced STING
expression in vivo. Tumor samples harvested from the immune-competent
mice at the endpoint were analyzed by SDS-PAGE and western blot with
indicated antibodies. One representative blot of n = 3 biological
replicates. Source data are provided as a Source Data file. For (B, C,
and F), data are presented as mean values ± SD. P-values were
calculated by unpaired Student’s t test with two-tailed analysis
without adjustments.
Tumor samples from the immune-competent mice were then analyzed to
determine the effect of birinapant on the tumor immune
microenvironment. Immunohistochemistry staining showed a significant
increase in the number of tumor-infiltrating CD8^+ T cells upon
birinapant treatment (Fig. [230]6D). This increase in
tumor-infiltrating CD8^+ T cells was confirmed using unbiased
single-cell mass cytometry profiling (Fig. [231]6E, F and Supplementary
Data [232]4). Significantly, birinapant treatment restored STING
expression in vivo, supporting the results from the in vitro studies
(Fig. [233]6G). Further, the depletion of CD8^+ T cells markedly
reduced the antitumor efficacy of birinapant (Fig. [234]S11F).
Altogether, these results demonstrate that targeting the aberrant
LKB1-cIAP1-JAK1 trimolecular complex, such as using an IAPi birinapant,
restores STING expression in LKB1-mut tumors and leads to enhanced
cytotoxic T cell infiltration and immune-dependent anti-tumor activity
in an immune-competent mouse model.
To further examine whether IAPi could enhance the therapeutic effect of
the immune checkpoint inhibitor in LKB1-mut tumors, we established a
CMT167 cell-based mouse model that recapitulated the immunotherapy
resistance phenotype of patients. We found that the LKB1-KO CMT167
cells showed significantly enhanced in vivo tumorigenicity capacity
than the LKB1-WT control in the immune-competent mice (Supplementary
Fig. [235]S11G). Moreover, the LKB1-KO CMT167 cells showed resistance
to anti-PD1 (200 µg/mouse) immunotherapy treatment (Fig. [236]S11G, H).
For potential translational studies, we selected a clinical-stage IAPi,
AT406 (Fig. [237]2D), for the following combination test with an
anti-PD1 agent. While treatment of an IAPi AT406 alone slightly slowed
down the growth of the LKB1-KO CMT167 tumor as compared to the DMSO
control, the combination of AT406 with anti-PD1 lead to a significant
synergistic anti-tumor effect (Supplementary Fig. [238]S11H). Further,
depletion of CD8^+, but not CD4^+, T cells significantly abolished the
anti-tumor activity of the AT406 and anti-PD1 combination treatment
(Supplementary Fig. [239]S11I), suggesting the underlying immune
killing effect and CD8^+ T cell-dependency.
LKB1 status controls the cIAP1-JAK1 interaction and the JAK1-STING signaling
The molecular and cellular functional studies, coupled with in vivo
data, revealed the IAP-STING regulatory axis as a potential
immune-dependent vulnerability created by the aberrant LKB1-cIAP1 PPI
in LKB1-mut cells. The action of IAPi appears to be associated with the
IFNγ-JAK-STAT signaling pathway. Therefore, we next examined the
potential molecular mechanism that bridges the LKB1-IAP PPI to the
JAK-STAT-STING axis with a cIAP1-directed PPI profiling assay.
From the binary interaction studies between cIAP1 and the IFNγ-JAK-STAT
pathway proteins with a robust TR-FRET PPI
assay^[240]38,[241]61–[242]63, we found that cIAP1 strongly interacted
with LKB1 and JAK1, respectively (Fig. [243]7A and Supplementary
Fig. [244]S12A), suggesting an intricate interplay among LKB1, cIAP1,
and JAK1. The binary cIAP1-JAK1 PPI was further confirmed by a
GST-beads affinity pulldown assay with the overexpressed proteins
(Fig. [245]7B) and a co-immunoprecipitation assay with the endogenous
proteins (Fig. [246]7C). These results support the physical interaction
of JAK1 with cIAP1 under physiologically relevant conditions.
Fig. 7. Identification of the LKB1-cIAP1-JAK1 trimolecular complex in shaping
IAP- and STING-dependency in LKB1-mut cells.
[247]Fig. 7
[248]Open in a new tab
A TR-FRET PPI signal between cIAP1 and IFNγ-JAK-STAT pathway proteins.
Data were presented as the average of n = 3 independent experiments. B
GST-PD confirmation of cIAP1-JAK1 PPI in HEK293T cells. C Endogenous
interaction of cIAP1-JAK1 with the co-IP assay in A549 cells. D, E
Mapping of JAK1-binding domain on cIAP1 in HEK293T cells co-expressing
GST-JAK1 with and VF-tagged cIAP1 full-length (FL), N-terminal
truncation (N), C-terminal truncation (C) or BIR domain truncations
(B1, B2 and B3). F Competitive binding between LKB1 and JAK1 with
cIAP1. G, H Immunoblot showing indicated proteins (G) and qPCR showing
JAK1 mRNA expression (H) (n = 3 independent experiments) in isogenic
LKB1-KD H1299 cells treated with 1 ng/mL IFNγ or PBS for 1 h. I
STAT-driven ISRE transcriptional activatity in HEK293T cells
transfected with ISRE-luc reporter plasmid with (+) or without (−)
LKB1-WT overexpression, and treated with 1 ng/mL IFNγ or PBS for 24 h
(n = 4 independent experiments). J, K JAK1 protein (J) and mRNA
expression (K) (n = 3 independent experiments) upon IAP inhibitor
treatment in A549 cells treated with IAP inhibitors (500 nM) as
indicated for 24 h. L JAK1 ubiquitination in A549 cells with 18 h
birinapant (500 nM) followed by 6-hour MG132 (20 μM) treatment. M
Immunoblot showing indicated proteins in A549 cells treated with IFNγ
(1 ng/mL) or PBS for 1 h with (+) or without 24 h pretreatment of
birinapant (500 nM). N, O IRF1 (N) and TAP1 (O) mRNA expression in A549
cells treated with birinapant (500 nM), IFNγ (1 ng/mL), or in
combination for 24 h (n = 3 independent experiments). Source data are
provided as a Source Data file. For (B–G, J, and L, M), data are
presented as one representative blot of n = 3 independent experiments.
For (H, K, and N, O), data are presented as mean values ± SD. P-values
were calculated by unpaired Student’s t test with two-tailed analysis
without adjustments.
To understand the structural basis of the interaction, truncation
studies were performed with cIAP1 domains (Fig. [249]1D). We found that
JAK1 was associated with the N-terminal domain of baculoviral IAP
repeats (BIR), particularly the BIR1-2 domain (Fig. [250]7D, E), which
is the same binding domain as LKB1 (Fig. [251]1E, F). These results
imply that LKB1 and JAK1 may compete for interaction with cIAP1. In
support of this notion, the level of JAK1 was decreased in the cIAP1
complex upon LKB1 overexpression (Fig. [252]7F). These results
confirmed the cIAP1-JAK1 interaction, which appears to be mutually
exclusive with the LKB1-cIAP1 interaction. A similar interaction mode
was detected between cIAP2 and JAK1 (Supplementary Fig. [253]S12B–D).
Due to the competitive nature of LKB1 and JAK1 for cIAP1 interaction,
the state of LKB1 interaction with cIAP1 may affect how cIAP1 regulates
JAK1 and the IAP dependency of cells. To test this, we utilized both
genetic and pharmacological perturbation approaches to examine the
effect of the LKB1 and IAP1 status on JAK1 and its signaling. First,
genetic knockdown of LKB1 led to a significant decrease in JAK1
protein, but not mRNA, suggesting a loss of JAK1 protein stability and
function (Fig. [254]7G, H). Supporting this result, the knockdown of
LKB1 impaired IFNγ signaling response, as shown with reduced STAT1
phosphorylation (Fig. [255]7G). Conversely, while overexpression of
LKB1-WT sensitized IFNγ signaling, as measured by the STAT
transcriptional reporter activity (Fig. [256]7I). Second, we found that
birinapant, which degrades cIAP1 and thus disrupts the cIAP1-JAK1 PPI,
rescued the expression of JAK1 protein, but not mRNA, in LKB1-mut cells
(Fig. [257]7J–L and Supplementary Fig. [258]S12E, F). This
birinapant-induced increase in JAK1 protein level is likely due to the
stabilization of JAK1 protein through inducing cIAP1 autodegradation
and disrupting cIAP1-JAK1 PPI (Fig. [259]7L and Supplementary
Fig. [260]S12G). Furthermore, the upregulation of JAK1 with the
birinapant treatment was correlated with significant downstream
signaling activation, as shown by increased IFNγ-induced STAT1/3
phosphorylation (Fig. [261]7M and Supplementary Fig. [262]S12H), STAT
transcriptional reporter activity (Supplementary Fig. [263]S12I, J),
and the expression of target genes IRF1 and TAP1 (Fig. [264]7N-O and
Supplementary Fig. [265]S12K, L). Further, we found that cIAP1 can be
tyrosine phosphorylated by JAK1, while treatment with a JAK1 inhibitor,
ruxolitinib, decreased cIAP1-JAK1 PPI and cIAP1 phosphorylation
(Supplementary Fig. [266]S12M). These results suggest that lung cancer
cells with LKB1 mutations have rewired cIAP1-JAK1 PPI, leading to
impaired IFNγ-JAK-STAT signaling.
These data suggest that in LKB1-WT cells, LKB1 sequesters cIAP1 from
the JAK1 complex, maintaining functional IFNγ-JAK-STAT signaling and
the basal level of STING. Loss of LKB1 in LKB1-mut cells permits cIAP1
to bind and degrade JAK1, impairing the JAK-STAT signaling, and leading
to STING downregulation. This IAP-induced JAK-STING pathway silencing
can be pharmacologically reversed by small molecule IAP inhibitors.
Discussion
Rewiring of tumor intrinsic circuitry enables cancer cell avoidance
from immune attacks, ultimately leading to therapeutic
failure^[267]10–[268]12. For example, lung adenocarcinomas harboring
mutations in LKB1 are highly aggressive, treatment-refractory, and
insensitive to immune checkpoint inhibitors^[269]9,[270]24–[271]26,
representing a major clinical challenge. It is urgent to develop
targeting strategies to enhance tumor response to therapies.
Altogether, the discovery of rewired LKB1-cIAP1 OncoImmune PPI
(Fig. [272]1), and the identification of IAP inhibitors as small
molecule immune response sensitizers in LKB1-mut LUAD cells
(Fig. [273]2), convergently point to a hypothesis that LKB1-mut tumors
with aberrant LKB1-cIAP1 PPI may develop IAP-dependency that can be
exploited as therapeutic target in an immune-dependent manner. This
LKB1-mediated OncoImmune PPI network further expanded our previously
identified LKB1 interactome landscape and suggested LKB1 as a major
oncogenic PPI hub protein^[274]38.
Previous studies have identified several potential therapeutic
vulnerabilities, such as HSP90, in LKB1 mutant lung cancer using cancer
cell line models, while the complex immune microenvironment remained to
be determined due to the use of cancer cell alone
culture^[275]28,[276]32,[277]64–[278]66. Our studies with the
OncoImmune PPI and HTiP approaches uncovered IAP and IAP-regulated
JAK1-STING innate immunity pathway as an immune-dependent vulnerability
in LKB1 mutated lung cancer cells for overcoming the immunotherapy
resistance. Mechanistic examination revealed an intricate dynamic
protein interaction hub by which LKB1 interacts with IAP in competition
with JAK1, which offers a working model that the loss of LKB1 in
LKB1-mut cancer cells allows IAP to engage JAK1 to control
JAK1-mediated IFNγ response and the effector STING pathway,
underpinning the IAP dependency. In support of this model,
pharmacological inhibition of IAP with birinapant synergizes with IFNγ
to reverse STING expression and function to promote tumor cytotoxic T
cell infiltration and induce apoptotic death of LKB1-mut lung cancer
cells. Importantly, such an immune-dependent mode of action of IAP
inhibitors in LKB1-mut cancer cells revealed by the HTiP approach is
strongly supported by in vivo data. Birinapant effectively induced
shrinkage of LKB1-mutant tumors in an immune-competent mouse model
while showing no effect in an immune-deficient nude mouse model,
presenting a potential therapeutic strategy for lung cancer patients
with mutated LKB1.
The regulatory complex uncovered in our study couples the status of IAP
to the activation state of the JAK1-STING pathway. Downregulation of
tumor intrinsic STING expression has been frequently observed in many
tumor types^[279]27,[280]28,[281]67–[282]70. However, the regulatory
mechanisms underlying STING expression control are context-dependent
and remain to be elucidated^[283]28,[284]71–[285]73. In the context of
LKB1-mut LUAD, STING expression downregulation has been associated with
transcriptional repression^[286]27,[287]28. Distinct from the
previously identified mechanism of epigenetic reprogramming mediated by
DNMT1 and EZH2^[288]28,[289]74, our data reveal a transcriptional
program by which the IFNγ-JAK-STAT immune response signaling pathway
dictates the STING expression in LKB1-mut cells. Our data further
suggest that tumor intrinsic STING expression could be a potential
biomarker for immune responsiveness, while restoring STING expression
could be effective therapeutic strategies in LKB1-mut LUAD. Our study
offers alternative approaches with the JAK1-directed mechanisms to
reactivate the STING pathway in various tumors, which may be
synergistic with the reported epigenetic modulation approach.
Dysregulation of IAP and JAK has been identified in a number of tumor
types. For example, overexpression of IAPs has been linked to tumor
progression, evasion of apoptosis, and poor prognosis^[290]75–[291]77.
On the other hand, the loss-of-function JAK mutations have been
associated with immune evasion and primary resistance to
immunotherapy^[292]13. In contrast to the aberrant IAP and JAK function
due to alterations at the genetic level, our data suggest that cIAP1
and JAK1 may be dysregulated at the protein level via rewired
protein-protein interactions. Our study suggests that sequestration of
IAP1 by LKB1 helps maintain a functional IFNγ-JAK-mediated immune
response in LKB1-WT cells, whereas, in LKB1-mut cells, IAP1 directly
impinges on JAK1 and impairs tumor immune response. Thus, IAP’s
pro-oncogenic activity could be enhanced by the loss of its bound LKB1
in LKB1 mutant cancer, while JAK1 could be directly weakened by IAP
interaction, leading to STING suppression and immune resistance. These
results implicate that such an immune response regulatory mechanism at
the JAK1 protein level might be a general phenomenon in other tumor
types with WT JAK1^[293]14. Moreover, the rewired JAK1 PPIs, such as
IAP1-JAK1 in LKB1-mut cells, might be a promising therapeutic target,
inhibition of which may sensitize tumors for enhanced immune response.
Like the demonstrated IAP inhibitors, the IAP/JAK1 PPI antagonists may
have the potential to specifically re-activate the STING activity for
therapeutic development.
The discovered IAP-JAK1 interaction suggests an emerging pathway for
the action of IAP inhibitors as anticancer immune enhancers. IAP
inhibitors, also known as SMAC mimetics, have been studied extensively
as potential anti-tumor agents^[294]78–[295]83. IAP inhibitors were
designed to block the inhibitory effect of IAP on caspase-mediated
apoptosis signaling and thus promote cell death. Our data suggest that
IAP inhibitors alone have minimal cell death-inducing effect on
LKB1-mut tumors. However, immune factors, such as IFNγ, is required for
IAP inhibitors to achieve their potent anti-tumor effect. IAP
inhibitors synergize with IFNγ to activate the RIPK1-dependent cell
death pathway in a colon cancer cell model. In LKB1-mut LUAD tumors,
however, the IAP inhibitor and IFNγ combination is JAK1-dependent
involving the STING activation. Thus, the effect of the IAP inhibitor
and IFNγ combination on its effector signaling could be
context-dependent, adding another layer of complexity to the emerging
immunomodulatory activities of IAP
inhibitors^[296]40,[297]47,[298]84–[299]88.
Our results may have significant implications for the treatment of
immune cold tumors. LKB1-mut LUAD has been characterized to exhibit an
immune suppressive phenotype with multiple immune-cold features, such
as low PD-L1 expression, low tumor mutational burden, altered
proinflammatory cytokine profiles, reprogrammed immune infiltration,
remodeled extracellular matrix, and recently demonstrated STING
expression downregulation^[300]9,[301]27–[302]29,[303]32–[304]36.
Therefore, effective strategies that re-inflame LKB1-mut tumors may
have significant therapeutic potential. Our data suggest that IAP
inhibitors may impact both immune infiltration and tumor-killing phase
in part through restoring tumoral STING expression, reinvigorating
STING-mediated DNA-sensing pathway, re-inducing innate immunity
cytokine and chemokine production, promoting chemotaxis of immune
cells, and re-sensitizing tumors to IFNγ-mediated immune
responsiveness. Our data suggest that IAP inhibitors may be promising
immunotherapy adjuvants with immune checkpoint inhibitors or STING
agonizts to further overcome LKB1-mut-associated immune resistance.
To further determine the translational potential of our findings,
future studies with long-term survival experiments in more
disease-relevant GEMM models are needed. In our in vivo study, the mice
in the control group reached the maximal tumor burden permitted in a
short-term within two weeks. Therefore, we evaluated the in vivo
efficacy of IAP inhibitors within the same short-term treatment, and
collected paired tumor samples to gain mechanistic insights into the
LKB1-IAP-JAK regulatory complex in LKB1 mutated tumors. However, we
recognize the importance of long-term survival studies to
comprehensively evaluate the in vivo efficacy of IAP inhibitors and its
combination effect with anti-PD1 immunotherapy for clinical
translation. In addition, we used subcutaneous syngeneic mouse models,
which may have distinct immunological profiles as compared to the
orthotopic GEMM models. Future studies using additional LKB1-mut LUAD
GEMM models by considering the clinically relevant co-occuring mutated
genes, such as KRAS and TP53, are needed to delve deeper into the
long-term effect and to determine potential treatment-associated
adaptive resistance mechanisms.
Methods
Ethics statement
This research complies with all relevant ethical regulations. All
animal studies were approved and conducted according to the Emory
University Institutional Animal Care and Use Committee (IACUC)
guidelines. A cumulative tumor burden scoring system was used for
comprehensive assessment of the animal’s condition related to tumor
development, including tumor size limit of 18–20 mm, body weight, tumor
ulceration, mobility and other body conditions. Human peripheral blood
mononuclear cells from healthy volunteer donors were purchased from
ATCC.
In vitro cell culture
All cell lines were incubated at 37 °C in humidified conditions with 5%
CO[2]. Human embryonic kidney 293 T cells (HEK293T; ATCC, CRL-3216)
were maintained in Dulbecco’s Modified Eagle’s Medium (DMEM; Corning,
#10-013-CV). Human non-small cell lung cell lines, including A549
(ATCC, CCL-185), Calu-1 (ATCC, HTB-54), H1299 (ATCC, CRL-1848), H1755
(ATCC, CRL-5892), H1792 (ATCC, CRL-5895), H1944 (ATCC, CRL-5907), H23
(ATCC, CRL-5800), H292 (ATCC, CRL-1848), and H460 (ATCC, HTB-177), were
cultured in Roswell Park Memorial Institute(RPMI) 1640 medium. The
isogenic LKB1-wildtype and mutant cells were generated by
lentivirus-transduction of corresponding shRNA or cDNA plasmids in the
parental cells. Human immune cells, including Jurkat T cell (ATCC,
TIB-152), NK92-MI (ATCC, CRL-2408), peripheral blood mononuclear cells
(normal PBMC, ATCC, PCS-800-011), CD8^+ T cells (Stemcell, 200-0164)
and CD56^+ NK cells (Stemcell, 70037) were cultured in RPMI-1640
medium. Cell culture medium was supplemented with 10% fetal bovine
serum (ATLANTA biologicals, #S11550) and 100 units/ml of
penicillin/streptomycin (Cell Gro, Cat# 30-002-CI). All cell lines were
authenticated via short tandem repeat (STR) profiling at regular
intervals to confirm their identity. We routinely tested all cell lines
for mycoplasma contamination using PCR-based MycoStrip® by InvivoGen
kit (Cat# rep-mysnc-100).
Cell line and mouse model for in vivo studies
Mouse lung adenocarcinoma cells, WRJ388^[305]59 and KW634 (generous
gifts from Dr. Kwok-Kin Wong)^[306]89 were cultured in RPMI-1640
medium. CMT167 cells (CancerTools, Cat. #: 151448) were cultured in a
DMEM medium. CMT167 cells were derived from metastasizing mouse lung
cancer CMT64 cells and carry KRAS^G12V mutation, and are TP53- and
LKB1-WT^[307]90–[308]93. CMT167 LKB1-WT scramble control and the
isogenic LKB1-knockout cells were generated by lentiviral-based
CRISPR/Cas9 technique with sg RNA targeting mouse LKB1 or non-target
control (Applied Biological Materials, cat# 45728114 and K010). Five-
to 6-week-old male athymic nude mice (18–20 g) were purchased from
Harlan Laboratories. Syngeneic immune-competent KL (Kras^G12D/Lkb1^-/-)
female mice were generated from Kras^G12DLkb1^fl/flRosa-luc GEMM
(KL-GEMM) model^[309]36,[310]59. Eight-week-old C57BL/J6 mice were
purchased from The Jackson Laboratory.
Plasmids
Plasmids for mammalian expression of Glutathione S-transferase- (GST),
Nanoluc (NLuc-), Venus-Flag- (VF), and Human influenza hemagglutinin-
(HA) tagged proteins were generated using the Gateway cloning system
(Invitrogen, Waltham, MA, USA) according to the manufacturer’s
protocol. The WT STK11 (LKB1) gene in pDONR223 was purchased from the
OpenBiosystem Kinome Entry ORF set. Other genes, such as
OncoImmune-related genes, in pDONR221, were purchased from Human
ORFeome V8.1 set. HA-Ubiquitin was purchased from Addgene
(#18712)^[311]94. The LKB1 and cIAP1 domain truncations in pDONR223
were generated using PCR and Gateway cloning system. The LKB1
mutations, including K78I and nonsense mutations, were introduced using
QuikChange Lightning Site-Directed Mutagenesis Kit (Agilent
Technologies, Cat# 210518). STING and LKB1 cDNA were cloned into pHAGE
(plasmid HIV-1 Alex Gustavo George Enhanced) lentiviral vector for
lentivirus packaging. Plasmids for shRNA knockdown were purchased from
the MISSON shRNA library (Sigma-Aldrich, see Supplementary
Data [312]5). All the plasmids were confirmed by sequencing.
Transfection
Polyethylenimine(PEI, Cat#23966) transfection reagent was used for
plasmid transfection in HEK293T cell. FuGene HD (Roche, Cat# E2920) was
used in a ratio of 3 μl to 1 μg DNA for transfection in other cancer
cells.
Onco-Immune PPI mapping
We used BRET^n technology in a miniaturized uHTS 1536-well plate-based
format to assess PPIs between LKB1 and tumor-intrinsic immune-response
pathway proteins in live cells. NLuc- and Venus-fusion proteins allow
streamlined monitoring of protein expressions simultaneously with BRET
signal detection in a simple add-and-read mode. Briefly, HEK293T cells
(1500 cells in 4 μl per well) were cultured in 1536-well plates at
37 ^oC before they were co-transfected in wells with Venus-tagged
OncoImmune-related genes in combination with NLuc-tagged STK11 genes
using Linear polyethylenimines (PEIs, Polysciences, 23966).
Transfection was performed by adding 1 μl mixture of PEI (30 ng/μl) and
DNA (10 ng/μl) to 4 μl cell culture, assisted by robotic operations
with the Biomek NX^P Lab Automation Workstation (Beckman Colter). BRET
saturation assay was performed by titration of DNA amount to achieve
various NLuc- and Venus-tagged gene combinations.
After incubation for 48 h, Nano-Glo® luciferase substrate furimazine
(Promega, N1120) was added to the cells directly. The donor
luminescence signal at 460 nm and acceptor emission signal at 535 nm
were measured immediately using an Envision Multilabel plate reader
(PerkinElmer). The BRET^n signal is expressed as the ratio of light
intensity measured at 535 nm over that at 460 nm. The specific BRET^n
signal for the interaction of two proteins is expressed as net BRET^n,
which is defined as the difference in BRET signal with the
co-expression of two proteins and expression of the negative control
NLuc-protein only.
The relative amount of NLuc-tagged protein expression was measured by
the luminescence signal at 460 nm (L460) during the BRET^n signal
measurement in the 1536-well white plate (Corning, 3727); while the
Venus acceptor protein expression was detected by the Venus
fluorescence intensity (FI) with excitation at 480 nm and emission at
535 nm in 1536-well black clear-bottom plate (Corning, 3893). Cells
were seeded and transfected side-by-side under the same conditions for
the 1536-well white plate for BRET^n measurement and black plate for
Venus FI measurement. The ratio of relative amount of acceptor over
donor protein expression (Acceptor/Donor) was defined as Venus FI/L460.
This intensity ratio should be proportional to the acceptor/donor
ratio.
The PPI signal was quantified by fitting BRET^n saturation curves,
including one curve for PPI and two curves for empty NLuc (ctrl1) and
Venus (ctrl2) controls, with each in four replicates. The saturation
curves are based on the equation
[MATH:
Y=BRE<
/mi>Tmax⋅XBRET50+X
:MATH]
1
where Y is net BRET^n, and X is Acceptor/Donor. The negative X and Y
values were excluded from the analysis. For quantitative analysis, the
area under the curve (AUC) was computed as a measurement of BRET^n
signal, as AUC integrates both the amplitude and shape of the
saturation curve^[313]95,[314]96. The fold-of-change (FOC) was
calculated as AUC[PPI]/AUC[max(ctrl1, ctrl2)] to express the difference
between PPI and empty vector controls, and statistical significance
(P[FOC]) was calculated using Student’s t-test to estimate the
likelihood that AUC[PPI] is different from the AUC[ctrl1&ctrl2].
Immune and tumor cell co-culture assay
PBMCs, primary CD8^+ T cells or CD56^+ NK cells as effector immune
cells, and various human lung adenocarcinoma cells as target cells were
used for co-culture assays. Tumor cells were seeded at a specific
density in 384-well cell culture plate (Corning #3764). Twenty-four
hours later, the effector were then thawed and co-cultured in RPMI-1640
medium with tumor cells in a dose-dependent manner for four days. CD3
monoclonal antibody (100 ng/mL, OKT3, ThermoFisher) and human
recombinant interleukin-2 (10 ng/mL, PeproTech) were used as activation
cocktail to activate immune cells to supply immune killing factors,
such as Perforin and Granzyme B.
Cell proliferation measurement
The co-culture assay plates were imaged using the IncuCyte® S3
Live-Cell Analysis System (Essen Biosciences). The cancer cell
proliferation was monitored and characterized as the percentage of
confluence using the IncuCyte® basic analysis module. Because of the
size distinction between effector immune cells and target cancer cells,
the area filter of > 400 μm^2 was used to select cancer cells that are
larger in size.
Cell viability measurement
Cell Titer Blue (Promega, G8081) was added to each well. The plates
were incubated for desired time at 37 ^oC to allow the generation of
sufficient signal within the linear range. The fluorescence intensity
of each well was read using an PHERAstar FSX multi-mode plate reader
(Ex 545 nm, Em 615 nm; BMG LABTECH). The cancer cell viability was
determined using Cell Titer Blue assay by subtracting the background
signal from the PBMC alone control.
High-throughput immunomodulator phenotypic (HTiP) screen
The primary HTiP screen was performed^[315]40. Briefly, parental H1755
cells harboring LKB1 mutation were seeded in 384-well cell culture
plate (2000 cells/well in 40 μl medium; Corning, Cat#3764) and
co-cultured with PBMCs (1000 cells/well in 10 μl media containing the
activation cocktail). The 2036 Emory Enriched Library (EEL) compounds
(100 nl) were used^[316]40–[317]42. Compounds were added into wells in
each plate using Biomek NXP Automated Workstation (Beckman) from a
compound stock plate to give the final concentration of 2 μM. A
parallel screening was performed with H1755 cells alone (2000
cells/well) in 50 μl medium containing the same amount of activation
cocktail in the absence of PBMCs. After 4 days of incubation,
image-based cell proliferation readouts followed by biochemical-based
cell viability measurements were used to examine the compound effect on
cancer cell growth. The percentage of control (%C) was calculated using
the equation 100X(S[compound]-S[blank])/(S[positive]-S[blank]) (2),
where S[positive] and S[blank] are the corresponding average of the
cell fluorescence intensity for wells with DMSO containing PBMCs/medium
only or plus cancer cells, respectively. The immune killing selectivity
index was calculated using the equation %C[-PBMC]/%C[+PBMC] (3).
Quantitative real-time polymerase chain reaction (qPCR)
Total RNA was isolated from cell lysates using E.Z.N.A.® Total RNA Kit
I (Omega, Cat# R6834-01) and digested with DNase I (Invitrogen, Cat#
18068-015). A total of 1 μg RNA was subjected to cDNA synthesis using
SuperScript™ III First-Strand Synthesis System(Invitrogen, Cat#
18080051) followed the manufacture’s instruction. Reverse transcribed
cDNA or isolated cytoplasmic DNA was diluted 1:5 ~ 1:10 in
nuclease-free water. qPCR was performed using SYBR Green Supermix
(Bio-rad, Cat# 1725272) in Mastercycler® RealPlex PCR System
(Eppendorf) with primers as listed in Supplementary Data [318]1. All
the primers were ordered from Eurofins Genomics LLC. (Louisville, KY,
USA). The following thermal cycling conditions were used for IFNβ:
50 °C for 2 min; 95 °C for 10 min; 40 thermal cycles (94 °C for 10 sec,
59 °C for 30 sec, 72 °C for 45 sec and 75 °C for 29 sec)^[319]97. For
other genes, the following thermal cycling conditions were used: 95 °C
for 2 min; 40 thermal cycles (95 °C for 15 sec, 60 °C for 15 sec and
72 °C for 20 sec). RNA expression was normalized to GAPDH expression.
Data were conducted a comparative analysis of relative expression by
2^-ΔΔCt method. Primer information of qRT-PCR was shown in
Supplementary Data [320]1.
Isolation of cytoplasmic dsDNA
Cytoplasmic DNA was extracted by using mitochondrial DNA isolation kit
(BioVision, Cat# K280-50) according to the modified manufacturer’s
instructions. Briefly, 0.5 × 10^6 cells were lysed with 1 × cytosol
extraction buffer, homogenized by dounce tissue grinder(50–70 times),
and then the nuclei and mitochondrial fractions were removed by
centrifugation according to the manufacturer’s instructions.
Cytoplasmic DNA from lysate was isolated by QIAamp DNA Mini Kit(QIAGEN,
Cat# 51304). Isolated cytoplasmic DNA was purified by RNaseA (Thermo
Fisher Scientific, Cat# EN0531) to remove RNA contamination. The amount
of mtDNA in the cytosol was determined by qRT-PCR using MT-ND1 primers.
The amount of nuclear DNA in the cytosol was determined by qRT-PCR
using three different sets of primers designed for different
chromosomes as described previously^[321]28. The sequences of the
primers are listed in Supplementary Data [322]1. 40 ng DNA per well
template was used for PCR analysis as described in qRT-PCR.
Transcriptome (RNA-seq) analysis
The birinapant-induced immune-dependent transcriptome change was
analyzed by mRNA sequencing service at Novogene Corporation Inc.
(Sacramento, CA,USA). Briefly, LKB1-mut cancer cells were treated by
birianpant with or without PBMC co-culture for 24 hours. Then immune
cells were removed by washing monolayer cancer cells three times with
1X PBS. The remaining surface-attached cancer cells were harvested for
RNA sample preparation. The total RNA was isolated using E.Z.N.A.®
Total RNA Kit I. RNA sequence reads were aligned to the human reference
genome (GRCh38). Significantly up- or downregulated differential
expression genes (DEGs) were identified using |log[2](Fold-Change
(FC)) | ≥ 1 and adjusted P-value ≤ 0.05. Pathway enrichment analysis
was performed using Metascape^[323]98.
Bioinformatics analysis
The Gene Set Enrichment Analysis (GSEA) analyses was performed as
described previously^[324]41. Briefly, the GSEABase package in R Studio
was used to score the indicating gene sets. The Hallmark gene sets
available from the Molecular Signatures database (MSigDB)^[325]99 were
used as the reference gene sets. The rank of genes in indicating
pathways was used in accordance with the birinapant-induced
immune-dependent DEGs, or DEGs identified between LKB1-WT and LKB1-mut
lung adenocarcinoma patient samples from the Genomic Data Commons (GDC)
Data Portal ([326]https://portal.gdc.cancer.gov/). The normalized
enrichment score (NES) was calculated to reflect the degree in which a
set of genes is overrepresented at the extremes (top or bottom) among
the entire ranked list. All GSEA analyses were performed strictly
according to the instructions
([327]https://www.bioconductor.org/packages/release/bioc/vignettes/GSEA
Base/inst/doc/GSEABase.pdf). As for statistical significance,
|NES | > 1 with a P-value and False discovery rate (FDR) < 0.05 was
considered as significantly enriched.
Cell lysate-based affinity pulldown assay
Cell lysate-based affinity-pulldown assay was performed as we
previously described^[328]38,[329]41,[330]61,[331]62,[332]100–[333]104.
Briefly, cell lysate were prepared in nonidet P-40 (NP-40) lysis buffer
(200 μL contains 1% NP-40 (IGEPAL CA-630, Sigma-Aldrich), 20 mM
Tris-HCl, 150 mM NaCl, 5% glycerol and 2 mM EDTA, supplemented with
protease inhibitor cocktail (Sigma-Aldrich, Cat# P8340), phosphatase
inhibitor cocktail 2 (Sigma-Aldrich, Cat# P5726), and phosphatase
cocktail 3 (Sigma-Aldrich, Cat# P0044)). Cell lysate were subjected to
affinity-based pulldown using glutathione-conjugated sepharose beads
(GE, Cat# 17-0756-05) for GST-pulldown, EZview Red Anti-Flag M2
Affinity Gel (Sigma-Aldrich, Cat# F2426) for flag-immunoprecipitation,
or EZview Red Anti-HA Affinity Gel (Sigma-Aldrich, Cat# E6779) for
HA-immunoprecipitation. Cell lysate were incubated with beads at 4 °C
for 2 h. Beads were washed with NP-40 lysis buffer for three times.
Pulldown or immunoprecipitated protein complex were eluted by boiling
the beads at 95 ^oC for 5 min in 2x Laemmli buffer (Bio-rad, Cat#
1610737, supplemented with 200 mM DL-Dithiothreitol (DTT)). Samples
were then analyzed by SDS-polyacrylamide gel electrophoresis and
immunoblotting with desired antibodies.
Co-immunoprecipitation (Co-IP) with endogenous proteins
The endogenous co-IP assay was performed as described
previously^[334]41,[335]62,[336]100. Briefly, cell lysate were prepared
in NP-40 lysis buffer. DTT (10 mM) and N-Ethylmaleimide (5 mM) were
added for cell lysate used for protein ubiquitination level
measurement. Cell lysates with total ~ 1.5 mg of proteins were used for
immunoprecipitation by incubating with desired protein antibody or IgG
control at 4 ^oC for 16 h. Then, protein A/D agarose beads were added
to the cell lysate and antibody mixture for incubation at 4 ^oC for
1 h. Beads were then washed with NP-40 lysis buffer three times.
Immunoprecipitated protein complex were eluted by boiling for 5 min at
95 ^oC in 2x Laemmli buffer without DTT. Then add DTT into the
supernatant at a final concentration of 20 mM and then boil for another
5 min at 95 ^oC. Samples were analyzed by SDS-polyacrylamide gel
electrophoresis and immunoblotting with desired antibodies.
Time resolved-fluorescence resonance eEnergy Transfer (TR-FRET) assay
TR-FRET assay was performed as previously described^[337]38. The FRET
buffer used throughout the assay contains 20 mM Tris-HCl, pH 7.0, 50 mM
NaCl, and 0.01% NP-40. Cell lysate from HEK293T cells expressing
GST-tagged and Venus-flag (VF) tagged donor and acceptor proteins were
prepared in 1% NP-40 buffer. Cell lysate were serially diluted in FRET
buffer and mixed with anti-GST-Terbium antibody (1:2000, Cisbio US Inc,
Cat# 61GSTTLB). The plate was centrifuged at 200xg for 5 min and
incubated at 4 ^oC for overnight. TR-FRET signals were measured using
the BMG Labtech PHERAstar FSX reader with the HTRF optic module
(excitation at 337 nm, emission A at 665 nm, emission B at 620 nm,
integration start at 50 μs, integration time for 150 μs and 8 flashes
per well). All FRET signals were expressed as a TR-FRET ratio: F665 nm
/F620 nm x 10^4.
Western blot (Immunoblot)
Proteins in the SDS sample buffer were resolved by 10% SDS
polyacrylamide gel electrophoresis (SDS-PAGE) and were transferred to
nitrocellulose filter membranes at 100 V for 2 h at 4 °C. After
blocking the membranes in 5% nonfat dry milk in 1 × TBST (20 mM
Tris-base, 150 mM NaCl, and 0.05% Tween 20) for 1 h at room
temperature, membranes were blotted with the indicated antibodies at
4 °C overnight. Membranes were washed by 1×TBST for three times, 15 min
each time. SuperSignal West Pico PLUS Chemiluminescent Substrate
(Thermo, #34580) and Dura Extended Duration Substrate (Thermo, #34076)
were used for developing membranes. The luminescence images were
captured using ChemiDoc™ Touch Imaging System (Bio-Rad).
Antibodies for western blot
The following antibodies were used for western blot: Flag-HRP (Sigma,
Cat# A8592, dilution 1:4000), GST-HRP (Sigma, Cat# A7340, dilution
1:5000), GST(Cell Signaling Technology, Cat.# 2624S, dilution 1:1000),
Flag(Cell Signaling Technology, Cat.# 14793, dilution 1:1000), HA (Cell
Signaling Technology, Cat.# 3724S, dilution 1:1000), β-actin (Sigma,
Cat.# A5441, 1:5000 dilution), STING(Cell Signaling Technology, Cat.#
13647, dilution 1:1000), TBK1(Cell Signaling Technology, Cat.# 3504,
dilution 1:1000), phospho-TBK1(S172) (Cell Signaling Technology, Cat.#
5483, dilution 1:1000), IRF3(Abcam, Cat.#ab68481, dilution 1:1000),
phosphor-IRF3(S386)(Abcam, Cat.# 76493, dilution 1:500),
LKB1(Invitrogen, Cat.# AHO1392, dilution 1:500), JAK1(Cell Signaling
Technology, Cat.# 50996S, dilution 1:1000), JAK2(Cell Signaling
Technology, Cat.# 3230S, dilution 1:1000), TYK2 anitbody(Cell Signaling
Technology, Cat.# 14193S, dilution 1:1000),STAT1(Cell Signaling
Technology, Cat.# 9172S, dilution 1:1000), phosphor-STAT1(Y701)(Cell
Signaling Technology, Cat.# 7649S, dilution 1:1000), STAT3(Cell
Signaling Technology, Cat.# 4904S, dilution 1:1000),
phosphor-STAT3(Y705)(Cell Signaling Technology, Cat.# 9145S, dilution
1:1000), cIAP1(Cell Signaling Technology, Cat.# 7065, dilution 1:1000),
cIAP2(Cell Signaling Technology, Cat.# 3130, dilution 1:1000),
XIAP(Cell Signaling Technology, Cat.# 14334S, dilution 1:1000),
Ubiquitin(Cell Signaling Technology, Cat.# 3933, dilution 1:1000),
anti-mouse IgG-HRP (Jackson ImmunoResearch, Cat.# 115-035-003, dilution
1:5000), and anti-rabbit IgG-HRP (Jackson ImmunoResearch, Cat.#
111-035-003, dilution 1:5000).
Generation of lentivirus
HEK293T cells (5 × 10^6) were seeded onto a 6-well plate and
transfected using PEI transfection reagent with 2 μg of
pHAGE-puro-lentivirus-based expression vector, or shRNA or sgRNA
vectors, together with 1.6 μg pCMV-dR8.91 and 0.66 μg of
pCMV-VSVG^[338]105. Forty-eight to 72 h after transfection, the
conditioned media containing lentivirus particles were collected and
centrifuged at 4 ^oC, 2500 × g for 15 min. Then media were filtered by
0.45 μm PVDF filter (Millipore, Cat# SLHV033RS) and stored at – 80 ^oC.
Interferon-stimulated response element (ISRE) and Interferon Gamma-Activated
Sequence (GAS) luciferase reporter assay
HEK293T cells transiently expressing GAS or ISRE-driven firefly
luciferase and internal control renilla luciferase or A549 cell stably
expressing GAS or ISRE-driven firefly luciferase were used for the
luciferase reporter assay. Cells were transfected with VF-tagged
plasmids or treated with birinapant and/or IFNγ as indidated. Renilla
and Firefly luciferase activities were measured by an Envision
Multilabel plate reader (PerkinElmer) using and Dual-Glo luciferase kit
(Promega, Cat# E2920) according to the manufacturer’s instructions. The
normalized luminescence was calculated as the ratio of luminescence of
Firefly luciferase over the luminescence of Renilla luciferase for
HEK293T cells. For stable A549 ISRE- or GAS-luciferase reporter cells,
only the firefly luciferase signal was measured.
Cytosolic double-strand DNA Staining
Cells were cultured on chambered cell culture slides (ibidi, Cat#
80826). Cells were treated with indicated treatments or vehicle control
for 24 hours. Following the treatment, cells were incubated with
culture media containing PicoGreen dsDNA stain (200-fold dilution)
(Thermo Fisher, Cat# P7581) for 1 h. Then cells were incubated with
RPMI1640 medium (without FBS) containing 200 nM MitoTracker™ Red
reagent (Thermo Fisher, Cat# M7512) for 20 min. Then, cells were fixed
with 4% paraformaldehyde in PBS (Fisher scientific, Cat# 50-980-487)
for 10 min. Cells were then washed twice with PBS and stained with
Hoechst 33342 (Thermo Fisher Scientific, Cat# H3570) for 10 min. The
image was acquired using Nikon A1R HD25 inverted confocal microscopes.
STING-HiBiT expression assay
STING-HiBiT-engineered cells were generated by essentially following
the protocol for CRISPR-mediated HiBiT tagging of endogenous proteins
(Promega)^[339]48. STING-HiBiT expression was monitored using the
Nano-Glo® HiBiT Lytic Detection System (Promega).
Apoptosis assay
The cell apoptosis assay was performed essentially by following
manufacturer’s protocol using Incucyte® Caspase-3/7 Green Dye
(SARTORIUS, Cat# 4440) (1:1000 dilution at final concentration of
5 mM). Data was aquired and analyzed using IncuCyte® S3 Live-Cell
Analysis System.
Transwell-based cell migration assay
Chemotaxis of immune cells was measured using a transwell-based cell
migration assay essentially by adapting previously described
methods^[340]106. Briefly, A549 Nuclight-red cells (SARTORIUS, #4491)
was plated at the bottom chamber of the transwell plate. Twenty-four
hours later, cells were treated with birinapant, JAK inhibitor, or
STING inhibitor as indicated followed by adding Jurkat T cells or
NK92-MI cells pre-labeld with CellTracker™ Green CMFDA Dye
(ThermoFisher, # C2925) at the upper chamber. Forty-eight hours later,
data was acquired and analyzed using the ImageXpress Micro HCS Imaging
System.
cGAMP ELISA assay
The cellular 2’,3’-cGAMP level was quantified using the 2’,3’-cGAMP
ELISA kit from Cayman Chemical (#501700). Briefly, A549 cells were
seeded in 10 cm plates and treated with either Birinapant 500 nM or
human IFNγ 1 ng/mL. Twenty-four hour after treatment, cells lysate were
prepared in PBS buffer supplemented with 1% NP-40. The cell lysate was
centrifuged at 10,000 × g, 4 ^oC for 10 min, and the cleared
supernatant were subjected to the analysis. The 2’,3’-cGAMP levels in
the lysate were measured by following the manufacturer’s instructions.
Animal studies
The WRJ388 and CMT167 cell-based transplantation mouse model was used
to evaluate the IAP inhibitor’s anti-tumor efficacy in vivo similarly
as described previously^[341]36,[342]59. Briefly, a total of 2×10^6
exponentially growing WRJ-388 cells were transplanted into syngeneic
immune-competent male KL mice through subcutaneous injection. CMT167
cells (1 × 10^6) were allografted subcutaneously in C57BL/J6 mice. Male
nude mice were used as immune-deficient control. Tumors were measured
every 3 days, and the formula for tumor volume is (length x width x
width)/2. Mice were randomly allocated into control and treatment
groups. Both Vehicle control (Captisol or DMSO), IAP inhibitors
(birinapant, intraperitoneal injections, 10 mg/kg once per 3 days; or
AT406, intraperitoneal injections, 30 mg/kg once per 3 days), anti-PD1
(InVivoMAb anti-mouse PD1, CD279, BioXcell, intraperitoneal injections,
200 µg/mouse once per 3 days), or combination treatments began when the
tumor kept growing between the 2 consecutive measuring and tumor volume
reached about 100–300 mm^3. Birinapant was dissolved in 12.5% Captisol
(MedChemExpress) in distilled water. AT406 was dissolved in PBS. The
solution was fresh diluted and sonicated before using. At the indicated
endpoint, tumor samples were harvested for further analysis. For CD8^+
T cell depletion, anti-mouse CD8a antibody or rat IgG2b isotype control
(Bioxcell, #BE0061 and #BE0090) was intraperitoneal injected at
10 mg/kg one day before birinapant treatment. All animal studies were
approved and conducted according to the Emory University Institutional
Animal Care and Use Committee (IACUC) guidelines. Mice were housed in a
pathogen-free animal facility under a 12-hour light/dark cycle with
constant ambient temperature (22 ± 2 °C) and humidity (40–60%). They
were provided ad libitum access to food and water. C57BL/6 J mice (Mus
musculus) were used in this study. The mice were obtained from The
Jackson Laboratory and were C57BL/6 J (000664). All mice were of 6–8
weeks old and maintained on a pure genetic background. The maximal
tumor size/burden was not exceeded according to the guidelines.
Single cell mass cytometry
Immune cell profiling was performed using single-cell mass
cytometry^[343]107–[344]110. Single-cell suspensions from mouse tumor
samples were stained with 16-marker panels using metal conjugated
antibodies according to manufacturer-suggested protocol using Mar
Maxpar® Mouse Sp/LN Phenotyping Panel Kit (Fluidigm, Cat# 201306).
Cells were fixed, permeabilized, and washed according to manufacturer’s
cell surface antigen staining protocol (Fluidigm). After antibody
staining, cells were incubated with intercalator solution, washed,
mixed with EQ Four Element Calibration Beads (catalog 201078), and
acquired with mass cytometer (all reagents from Fluidigm). Gating and
data analysis were performed with Cytobank
([345]https://www.cytobank.org/). Intact viable cells were identified
using cisplatin intercalator according to manufacturer-suggested
concentrations (Fluidigm). viSNE analysis was performed with Cytobank.
IHC analysis
Five-micron-thick paraformaldehyde-fixed OCT-embedded mouse lung
sections or formalin-fixed paraffin-embedded mouse tumor lung sections
were used for IHC analyses^[346]59. Slides were stained with primary
CD8 antibody (Abcam, Cat# ab237723), and horse anti-rabbit IgG (Vector,
Cat# MP-7401) was used as the secondary antibody. DAB substrate kit was
used to develop IHC signals. These samples were blinded and analyzed by
a lung cancer pathologist (G.L. Sica).
Statistics and reproducibility
All analyses were performed using GraphPad Prism version7.0 (GraphPad
Software, La Jolla, CA) or Microsoft Excel. The dose-dependent
PBMC-induced or small molecule induced cancer cell growth inhibition
curve was established using GraphPad Prism based on the Sigmoidal
dose-response (variable slope) fitting. Statistical significance was
assessed using student’s t test, or Wilcoxin test as indicated.
Statistical tests with exact p-values derived, the number of
independent biological replicates, and where applicable number of
analyzed cells are provided in the figures and figure legends.
P-values ≤0.05 were considered statistically significant. No
statistical method was used to predetermine the sample size. No data
were excluded from the analyses, the experiments were not randomized,
and the investigators were not blinded to allocation during experiments
and outcome assessment.
Reporting summary
Further information on research design is available in the [347]Nature
Portfolio Reporting Summary linked to this article.
Supplementary information
[348]Supplementary Information^ (4MB, pdf)
[349]41467_2025_57297_MOESM2_ESM.pdf^ (104.5KB, pdf)
Description of Additional Supplementary Files
[350]Supplementary Data 1^ (11.9KB, xlsx)
[351]Supplementary Data 2^ (87.1KB, xlsx)
[352]Supplementary Data 3^ (372.2KB, xlsx)
[353]Supplementary Data 4^ (9.2KB, xlsx)
[354]Supplementary Data 5^ (14.2KB, xlsx)
[355]Reporting Summary^ (2.4MB, pdf)
[356]Transparent Peer Review file^ (474.5KB, pdf)
Source data
[357]Source Data^ (6.7MB, xlsx)
Acknowledgements