Abstract

   Albumin-bound paclitaxel (nab-PTX) is an important chemotherapeutic
   drug used for the treatment of advanced and metastatic non-small cell
   lung cancer (NSCLC). One critical issue in its clinical application is
   the development of resistance; thus, a deeper understanding of the
   mechanisms underlying the primary resistance to nab-PTX is expected to
   help to develop effective therapeutic strategies to overcome
   resistance. In this study, we made an unexpected discovery that NSCLC
   with wild-type (WT) Liver kinase B1 (LKB1), an important tumor
   suppressor and upstream kinase of AMP-activated protein kinase (AMPK),
   is more resistant to nab-PTX than NSCLC with mutant LKB1.
   Mechanistically, LKB1 status does not alter the intracellular
   concentration of nab-PTX or affect its canonical pharmacological action
   in promoting microtubule polymerization. Instead, we found that LKB1
   mediates AMPK activation, leading to increased expression of SLC7A11, a
   key amino acid transporter and intracellular level of glutathione
   (GSH), which then attenuates the production of reactive oxygen species
   (ROS) and apoptotic cell death induced by nab-PTX. On the other hand,
   genetic or pharmacological inhibition of AMPK in LKB1-WT NSCLC reduces
   the expression of SLC7A11 and intracellular GSH, increases ROS level,
   and eventually promotes the apoptotic cell death induced by nab-PTX in
   vitro. Consistently, the combination of nab-PTX with an AMPK inhibitor
   exhibits a greater therapeutic efficacy in LKB1-WT NSCLC using
   xenograft models in vivo. Taken together, our data reveal a novel role
   of LKB1-AMPK-SLC7A11-GSH signaling pathway in the primary resistance to
   nab-PTX, and provide a therapeutic strategy for the treatment of
   LKB1-WT NSCLC by targeting the LKB1-AMPK-SLC7A11-GSH pathway.

   Keywords: Albumin-bound paclitaxel, LKB1, Non-small lung cancer, AMPK,
   SLC7A11, ROS

1. Introduction

   Lung cancer, as a highly heterogeneous disease, is one of the most
   common carcinomas and the leading cause of cancer-related death in the
   world [[41]1]. Histologically, lung cancer can be divided into small
   cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC) [[42]2]
   with the latter accounting for approximately 85 % of all lung cancers
   [[43]2]. Previously, the management of lung cancer was anchored to
   chemotherapy, radiotherapy and surgery operation [[44]2,[45]3]. With
   the advancement of lung cancer research, an increasing number of
   genetic alterations have been discovered, including tumor driver genes
   (EGFP, KRAS, ROS1) and immune checkpoint proteins (PD-L1), which
   significantly promote the development of targeted therapy and
   immunotherapy in NSCLC [[46]2,[47]4]. Despite the significant
   advantages of targeted therapy and immunotherapy in improving the
   overall prognosis of the patients with NSCLC, there are several
   important limitations, such as toxicity, drug resistance, and high cost
   [[48]5]. Meanwhile, targeted therapy and immunotherapy are typically
   used in combination with chemotherapy in clinical settings
   [[49]6,[50]7]. Therefore, chemotherapeutic drugs still play an
   important role in the treatment of lung cancer, especially metastatic
   lung cancer [[51]2,[52]8,[53]9]. Among these, albumin-bound paclitaxel
   (nab-PTX) is a widely used chemotherapeutic drug for the treatment of
   advanced and metastatic NSCLC [[54]2,[55]8,[56]9].

   Paclitaxel (PTX), a microtubule-binding drug, is well known to bind to
   β-tubulin to stabilize the microtubules of cells [[57]10], which blocks
   the dynamics of microtubules [[58]11] and induces apoptosis in cancer
   cells [[59]12,[60]13]. Nab-PTX is a novel formulation in which PTX
   binds to albumin nanoparticle through a non-covalent interaction
   [[61]14,[62]15]. By using albumin as an alternative vehicle for PTX,
   the allergic reactions induced by CrEL, a solvent to maintain the
   solubility and stability of PTX, can be avoided, and the efficacy of
   PTX could be increased for the accumulation of nab-PTX around the tumor
   by albumin [[63]16]. Currently, nab-PTX is approved by the U.S. Food
   and Drug Administration (FDA) for the treatment of metastatic breast
   cancer, locally advanced or metastatic non-small cell lung cancer, and
   advanced pancreatic cancer [[64]16]. The therapeutic efficacy of
   nab-PTX has been widely reported [[65]17,[66]18]. Nevertheless, primary
   or acquired resistance to nab-PTX is a major issue that limits its
   clinical application. Multiple mechanisms underlying this resistance
   have been reported, including the upregulation of Mcl-1 [[67]19],
   sustained induction of c-MYC [[68]20], and overexpression of ABCB1
   [[69]21].

   Liver kinase B1 (LKB1), a serine/threonine kinase, is a highly mutated
   gene in NSCLC, with a mutation rate of approximately 15 % [[70]22].
   LKB1 is usually regarded as a tumor suppressor: its loss-of-function
   mutation is responsible for the inherited tumor disorder Peutz-Jeghers
   syndrome (PJS), in which polyps form in the colon and the persons with
   PJS have a high risk for developing digestive cancer [[71]23]. The main
   physiological function of LKB1 is to directly phosphorylate and
   activate AMP-activated protein kinase (AMPK) at the Thr-172 residue of
   AMPK α-catalytic subunit [[72]24]. AMPK is probably one of the most
   important kinases in the control of metabolism, and activated AMPK is
   capable of promoting catabolism and inhibiting anabolic processes via
   direct phosphorylation of an array of key target proteins such as
   mammalian target of rapamycin complex 1 (mTORC1) and acetyl-CoA
   carboxylase (ACC) [[73]24]. As a downstream effector of LKB1, AMPK is
   presumably also a tumor suppressor, and activation of AMPK has been
   considered a potential therapeutic strategy for cancer [[74]24].
   However, intriguingly, some studies have reported that the activation
   of LKB1-AMPK signaling pathway protects the cells against multiple
   stimulus including energy starvation, nutrition deprivation, hypoxia
   and oxidative stress [[75][25], [76][26], [77][27], [78][28], [79][29],
   [80][30], [81][31], [82][32]]. Therefore, the suppression of the
   LKB1-AMPK pathway may serve as a potential therapeutic strategy for
   cancer.

   In this study, we demonstrate that the LKB1-AMPK-SLC7A11-GSH signaling
   pathway plays an important role in determining the therapeutic efficacy
   of nab-PTX: In LKB1-WT NSCLC cells, activation of LKB1-AMPK signaling
   promotes the expression of SLC7A11, increases the intracellular GSH
   level, reduces intracellular ROS level, and thereby impairs the
   therapeutic efficacy of nab-PTX. Suppression of AMPK significantly
   sensitizes LKB1-WT NSCLC cells to cell death induced by nab-PTX both in
   vitro and in vivo. Thus, our study provides experimental evidence that
   targeting the LKB1-AMPK-SLC7A11-GSH pathway could be a novel
   therapeutic regimen to overcome the primary resistance to nab-PTX in
   LKB1-WT NSCLC.

2. Materials and methods

2.1. Cells and cell culture

   H1299, A549, H460, H1975, H1650, HCC827, H23, H358 and HEK-293T cells
   were obtained from American Tissue Culture Collection (USA). Except
   HEK-293T cells, all cell lines were cultured in RPMI (Gibco, 11875093)
   supplemented with 10 % FBS (Gibco, 10099141C) and 1 % antibiotics
   (Penicillin/Streptomycin, Gibco, 15140122). HEK-293T cells were
   cultured in DMEM (Gibco, 11965092) supplemented with 10 % FBS (Gibco,
   10099141C) and 1 % antibiotics (Penicillin/Streptomycin, Gibco,
   15140122).

2.2. Cell viability measurement

   The cells were seeded at a density of 1000 cells/well in 96-well plates
   and incubated overnight. Following treatment, cell viability was
   measured using a Cell Counting Kit-8 (Dojindo, CK04) according to the
   manufacturer's protocol, and the IC[50] was calculated using GraphPad
   Prism 7 software.

2.3. Immunoblotting assay

   The cells were washed twice with cold PBS, scraped off in RIPA buffer
   (containing 1 % PMSF), and incubated on ice for 30 min. The cell
   lysates were centrifuged at 14,000×g for 15 min, and the supernatants
   were stored at −80 °C until further analysis. Protein concentrations
   were measured using a bicinchoninic acid (BCA) protein assay (Bio-Rad,
   5000201). Proteins were denatured by boiling and separated using 12 %
   SDS-PAGE. Proteins were transferred to PVDF membranes, which were
   activated in methanol. Membranes were immunoblotted with primary
   antibodies, followed by incubation with HRP-conjugated secondary
   anti-rabbit (AlpVHHs, 025-102-005) and anti-mouse (AlpVHHs,
   001-101-001) antibodies. The membranes were incubated with Western
   Chemiluminescent HRP Substrate (Millipore, WBLUF0500) for 1 min and
   then detected using a ChemiDoc™ Touch Imaging System (Bio-Rad). All
   antibodies were used at a 1:1000 dilution in 5 % bovine serum albumin
   (BSA) in TBST buffer for immunoblotting. The primary antibodies for
   this study are listed in the Antibody section of supplementary
   materials.

2.4. Immunofluorescence staining

   H1299 (sgC and sgLKB1) cells and A549 (EV, LKB1-WT, LKB1-KD) cells were
   seeded into 12-well plates with coverslips in each well and cultured
   overnight. Following treatment, the cells were incubated with 5 % BSA
   for 1 h to suppress nonspecific binding after washing and fixing with
   4 % PFA for 30 min at room temperature. Cells were subsequently
   incubated with anti-rabbit β-tubulin primary antibody (1:100) (Cell
   Signaling, 2118S) at 4 °C overnight, followed by incubation with goat
   anti-rabbit IgG (H + L) conjugated to Alexa Fluor^FM Plus 488 secondary
   antibody (1:300) (Thermo, A32731) for 1 h at room temperature. The
   stained cells were mounted with mounting medium containing DAPI for
   5 min and visualized under a confocal microscope (Zeiss LSM710). The
   β-tubulin channel (green) and DAPI (blue) were photographed and merged
   using ZEN Microscopy software.

2.5. Cell death measurement

   Cells were seeded in 6-well plates and cultured overnight. Following
   treatment, cell death was measured using an apoptosis kit (Bestbio,
   BB-4101-100T), according to the manufacturer's protocol. Briefly, the
   cells were harvested and washed carefully with cold PBS for 1 time.
   Annexin V-binding solution (400 μL) and 3 μL FITC dye were added, and
   the cells were incubated for 15 min on ice, followed by the addition of
   5 μL PI dye and 1 min of incubation. The cells were immediately
   analyzed by flow cytometry (Cytoexpert, Beckman Coulter).

2.6. Cell cycle measurement

   Cells were seeded in 6-well plates and cultured overnight. After
   treatment, the cells were collected, washed carefully with PBS, and
   fixed with 75 % ethanol at −20 °C. After 24 h of fixation, the cells
   were washed with PBS and stained with propidium iodide (PI) using the
   Cell Cycle Analysis Kit (Best Bo, [83]BB-4104) for 20 min. The cell
   cycle features were analyzed using a Cytoflex flow cytometer
   (Cytoexpert, Beckman Coulter). Data was analyzed using FlowJo 7.6
   software.

2.7. Immunohistochemical assay

   All xenograft tumors were fixed with 4 % PFA (Beyotime Biotechnology)
   overnight at 4 °C and then embedded in paraffin. The samples were
   subsequently sectioned into thin slices and mounted on slides, followed
   by deparaffinization in xylene and rehydration using a series of
   ethanol-water solutions. Antigen retrieval was performed by immersing
   the sections in citrate–acid buffer and heating them in a microwave
   oven. The slides were then blocked with 3 % hydrogen peroxide to block
   nonspecific activity. After rinsing, the slides were blocked with 5 %
   BSA and then incubated with Ki-67 (Cell Signaling, 9449T), p-ACC (Cell
   Signaling, 11818s), and p-H2A.X (Cell Signaling, 9718S) antibodies
   overnight at 4 °C. An immunohistochemical staining kit (BOSTER
   Biological Technology, Wuhan, China) was used for development.
   Microscopic images (Nikon, Japan) were captured and analyzed by
   experienced pathologists.

2.8. Gene expression profiling

   Gene mutation data for the LUAD cases were obtained using cBioPortal
   ([84]https://www.cbioportal.org/). Gene expression data for H1650 cells
   ([85]GSE69747) were obtained from the Gene Expression Omnibus (GEO)
   database ([86]https://www.ncbi.nlm.nih.gov/geo/). RNA sequencing of
   A549 and H1299 cells was performed by Novogene. Differentially
   expressed gene (DEGs) analysis was performed using the limma package.
   Gene set enrichment analysis (GSEA) was performed using GSEA software
   (Broad Institute).

2.9. ROS detection

   The cells were seeded in a 12-well plate and grown overnight before
   drug treatment. After treatment, CellROX Oxidative Stress Reagents
   (Thermo Fisher Scientific, [87]C10444) were added to the cells at a
   final concentration of 5 μM and incubated for 30 min at 37 °C. After
   incubation, the cells were harvested through trypsinization. The cells
   were centrifuged at 300×g for 5 min at 4 °C and washed twice with cold
   PBS. The cells were resuspended in 300 μL cold PBS and subjected to
   flow cytometry (Cytoexpert, Beckman Coulter) for ROS detection. Data
   was analyzed using FlowJo_V10 software.

2.10. Glutathione measurements

   Cells were seeded at a density of 1 × 10^4 cells per well in a 96-well
   plate (Corning, CLS3610) and grown overnight before the measurement of
   glutathione. Cells were seeded at a density of 5 × 10^3 cells per well
   in a 96-well plate (Corning, CLS3610) and grown overnight before
   treating with vehicle or 20 μM DL-Buthionine-(S,R)-sulfoximine
   hydrochloride (BSO, MCE, HY-106376B) or 40 μM EX229 (MCE, HY-112769)
   for 24 h. Glutathione levels were quantified using the GSH-Glo™
   Glutathione Assay kit (Promega, V6912) according to manufacturer's
   instructions. Total levels of glutathione were normalized to protein
   concentrations as determined by BCA assay.

2.11. RNA extraction and real Time-PCR (RT-PCR)

   Cells were seeded in 6-well plates and cultured overnight. According to
   the instructions of the manufacturer, total RNA was extracted from
   cultured cells using TRIzol™ Reagent (Invitrogen, 15596026). Total RNA
   was reverse-transcribed into cDNA using the iScript™ cDNA Synthesis
   Kits (BIO-RAD, 1708891). RT-PCR was performed with SsoAdvanced
   Universal SYBR® Green Supermix (BIO-RAD, 1725274) using standard
   conditions on the 7500 Fast Real-Time PCR system (Life Technologies).
   GAPDH was served as an internal control for normalization of target
   mRNA. Relative qPCR primers are shown in Supplementary Materials.

2.12. Plasmid constructions

   The CA-AMPK construct was a gift from Prof. Gang Li (addgene, #27632).
   The CA-AMPK sequence was cloned into the pCDH-CMV plasmid (addgene,
   #72265) using canonical digestion and ligation assay.
   pLV3-CMV-SLC7A11(human)-FLAG-Puro (Miaolingbio, [88]P46407) was used
   for the overexpression of SLC7A11.
   pLVX-U6-SLC7A11(human)-shRNA1-PGK-EGFP-Puro (Miaolingbio, [89]P33198)
   and pLVX-U6-SLC7A11(human)-shRNA2-PGK-EGFP-Puro (Miaolingbio,
   [90]P33255) were used for the knockdown of SLC7A11.

2.13. CRISPR–Cas9-mediated gene knockout

   Knockout of AMPKα1/α2 in human H1299 cells was performed using single
   guide RNAs (sgRNAs) and CRISPR–Cas9 technology. pLenti-Puro-sgAMPKα1
   (Miaolingbio, [91]P21236) and pLenti-Puro-sgAMPKα2 (Miaolingbio,
   P21235) were transfected into HEK293T cells with the psPAX2 packaging
   plasmid (addgene, #12260) and pVSV-G envelope-expressing plasmid
   (addgene, #138479) by using PEI MAX® (Polysciences, 24765). Supernatant
   (lentivirus) was collected and stored at −80 °C until ready to use.
   H1299 cells were infected with lentivirus for 48 h and selected with
   2 μg/ml puromycin (Sigma, P9620) for 5 days, and then single cells were
   sorted into 96-well plates. Single cells were maintained in RPMI with
   10 % (v/v) FBS and 1 % (v/v) penicillin/streptomycin at 37 °C in an
   incubator with 5 % CO2 for 3–4 weeks, and each colony was verified by
   Immunoblotting.

2.14. Tumor xenograft experiments

   Male BALB/c nude mice (4–5 weeks) were purchased from the Animal
   Facility of the University of Macau. Cancer cells (5 × 10^5 cells in
   180 μL of PBS) were subcutaneously inoculated into the flanks of BALB/c
   nude mice. Tumor volumes were measured by calipers every other day and
   calculated using the following formula: a^2 × b × 0.5, where a
   represents the smallest diameter and b is the diameter perpendicular to
   a. Mice were randomly split into two groups receiving either vehicle or
   nab-PTX (10 mg/kg) combined with or without BAY-3827 (10 mg/kg) every
   other day for 2 weeks. After the drug treatment period, mice were
   euthanized by using CO[2] asphyxiation. The xenografts were dissected,
   weighed, and preserved in a −80 °C freezer for immunohistochemical
   analysis of cell proliferation and cell death. Animal study was
   approved by the Ethics Committee of University of Macau (Approved
   protocol ID: UMARE-009-2022).

2.15. Statistical analysis

   All experiments were independently repeated more than two times with
   similar results. The statistical analyses were performed with GraphPad
   Prism 7 software. p values are indicated in figures and figure legends.
   P < 0.05 was considered statistically significant.

3. Results

3.1. LKB1 attenuates the therapeutic potency of nab-PTX in NSCLC

   A recent study reported that Kras-mutant NSCLC cells are more sensitive
   to the treatment of nab-PTX caused by the enhanced uptake of nab-PTX by
   activating the MEF/ERK signaling pathway [[92]33]. Co-occurrence of
   KRAS and LKB1 mutations is a common molecular feature in NSCLC, as
   summarized in [93]Table S1. Therefore, we deduced that NSCLC with
   mutant LKB1 might be more vulnerable to nab-PTX treatment. To verify
   our hypothesis, we performed the CCK-8 assay to determine the IC[50]
   value of nab-PTX in a panel of NSCLC cell lines. The results showed
   that LKB1-mutant NSCLC cells were generally more sensitive to nab-PTX
   than those NSCLC cells with wide type (WT) LKB1 ([94]Fig. 1A and B).
   The results of the cell death assay also showed that nab-PTX displayed
   more potent cytotoxicity in LKB1-mutant NSCLC cells ([95]Fig. 1C). To
   establish a causal relationship between LKB1 and nab-PTX cytotoxicity,
   we manipulated LKB1 status in NSCLC cells. First, deletion of LKB1 in
   H1299 cells (a cell line with WT-LKB1) using CRISPR/Cas9 technology
   ([96]Fig. 1D) significantly increased the efficacy of nab-PTX ([97]Fig.
   1E and F). Conversely, overexpression of WT-LKB1, but not its
   kinase-dead (KD) form, in A549 (a cell line with mutant LKB1) cells
   ([98]Fig. 1G) obviously reduced the cell death induced by nab-PTX
   ([99]Fig. 1H and I).

Fig. 1.

   [100]Fig. 1
   [101]Open in a new tab

   LKB1 impairs the apoptotic cell death induced by nab-PTX in NSCLC.

   A. The genetic alteration of LKB1 in H460, H23 and A549 cells, and an
   immunoblot showing the level of LKB1 in various NSCLC cell lines. B.
   The IC[50] values of nab-PTX in various NSCLC cell lines. C. Cell death
   measurement in various NSCLC cell lines treated with indicated
   concentrations of nab-PTX for 36 h. D. An immunoblot showing the
   deletion of LKB1 in H1299. E. Cell viability measurement in indicated
   H1299 cells treated with indicated concentrations of nab-PTX for 72 h.
   F. Cell death measurement in indicated H1299 cells treated with
   indicated concentrations of nab-PTX for 36 h. G. An immunoblot showing
   the levels of LKB1 in A549 cells as indicated. H. Cell viability
   measurement in indicated A549 cells treated with indicated
   concentrations of nab-PTX for 72 h. I. Cell death measurement in
   indicated A549 cells treated with indicated concentrations of nab-PTX
   for 36 h. J-K. Cell death measurement in indicated H1299 cells (J) and
   in indicated A549 cells (K) treated with 50 nM nab-PTX combined with or
   without 20 μM Z-VAD-FMK for 36 h. L. Images of the indicated H1299
   xenografted tumors at the endpoint of treatment with or without
   nab-PTX. M. Mass of the indicated H1299 xenografted tumors as shown in
   L. N. Images of the indicated A549 xenografted tumors at the endpoint
   of treatment with or without nab-PTX. O. Mass of the indicated A549
   xenografted tumors as shown in N. sgC, sgCtrl; sgL, sgLKB1; EV, empty
   vector; WT, wide type; KD, kinase dead. ns, no statistical
   significance; ∗∗p < 0.01; ∗∗∗p < 0.001. The mean ± s.d. are shown;
   n ≥ 3 independent experiments. Statistical analysis was performed using
   an unpaired, two-tailed t-test.

   Since PTX is responsible for the killing effect of nab-PTX [[102]15],
   we also examined the correlation between LKB1 status and PTX in NSCLC
   cells. Similar to the observation of nab-PTX, LKB1 deficiency enhanced
   PTX-induced cell death in H1299 cells ([103]Fig. S1A), while the
   reconstitution of WT-LKB1 impaired its killing effect in A549 cells
   ([104]Fig. S1B). Given that PTX can induce several types of cell death
   under distinct circumstances [[105][34], [106][35], [107][36],
   [108][37]], we further explored the type of cell death induced by
   nab-PTX in NSCLC cells. As showed in [109]Figs. S1C and S1D, Z-VAD-FMK
   (a pan-caspase inhibitor), but not ferrostatin-1 (Fer-1, a ferroptosis
   inhibitor) and GSK872 (a necroptosis inhibitor), significantly
   inhibited cell death induced by nab-PTX in both H1299 and A549 cells,
   indicating that nab-PTX induced apoptotic cell death in NSCLC cells.
   Meanwhile, we found that the addition of Z-VAD-FMK could also reduce
   the cell death induced by nab-PTX in LKB1 knockout (KO) H1299 cells
   ([110]Fig. 1J), accompanied by the downregulation of cleaved-caspase 7
   level ([111]Fig. S1E). Similar observations were made in A549 cells
   ([112]Fig. 1K, [113]Fig. S1F), suggesting that LKB1 status does not
   alter the type of cell death induced by nab-PTX. Finally, we tested
   whether LKB1 impairs the potency of nab-PTX in vivo using a nude mouse
   xenograft model. The results showed that nab-PTX displayed a stronger
   repression effect in the growth of LKB1-KO H1299 xenografts ([114]Fig.
   1L and M). Consistently, overexpression of LKB1 impaired the efficacy
   of nab-PTX in the growth suppression of A549 xenografts ([115]Fig. 1N
   and O). Taken together, these data suggest that LKB1 impairs the
   therapeutic potential of nab-PTX in NSCLC by protecting against cell
   death induced by nab-PTX.

3.2. LKB1 does not affect the uptake of nab-PTX and its canonical
pharmacological effects

   An earlier report showed that the higher susceptibility of Kras-mutant
   cells to nab-PTX is due to the enhanced uptake of nab-PTX [[116]33]. To
   explore whether LKB1 could also affect the uptake of nab-PTX in NSCLC
   cells, we performed liquid chromatography with tandem mass spectrometry
   (LC-MS/MS) to directly detect the intracellular concentration of PTX.
   The results showed that neither the deletion of LKB1 in H1299 cells nor
   the overexpression of WT-LKB1 in A549 cells altered the intracellular
   concentration of nab-PTX, compared to their respective parental cells
   ([117]Fig. 2A and B). Thus, our data indicated that LKB1 status has no
   direct role in regulating the uptake of nab-PTX. It is well known that
   the pharmacologic activity of PTX is via its direct binding to
   β-tubulin to promote the polymerization of microtubule, then leading to
   the stabilization and dysfunction of microtubule and subsequent cell
   cycle arrest at G2/M phase [[118]38]. To investigate whether LKB1 can
   affect the pharmacological effects of nab-PTX, we detected changes in
   microtubules and found that LKB1 status did not significantly alter the
   dynamics of microtubules induced by nab-PTX ([119]Fig. 2C and D). A
   cell cycle assay was also conducted to explore whether LKB1 impaired
   cell cycle profiles following treatment with nab-PTX. The results
   showed that nab-PTX induced a striking G2/M phase arrest in both H1299
   and A549 cells in a dose-dependent manner ([120]Fig. 2E). However, no
   significant change in cell cycle distribution was observed after the
   deletion of LKB1 in H1299 cells or after the overexpression of LKB1 in
   A549 cells ([121]Fig. 2F and G). Together, these data suggest that LKB1
   status does not affect the cellular uptake and the canonical
   pharmacological actions of nab-PTX.

Fig. 2.

   [122]Fig. 2
   [123]Open in a new tab

   LKB1 does not affect the uptake of nab-PTX and its canonical
   pharmacological effects

   A-B. Intracellular concentration measurement of PTX in indicated
   H1299 cells (A) and in indicated A549 cells (B) upon treatment with
   50 nM nab-PTX for 6 h. C-D. Confocal microscopy images of microtubule
   distribution in indicated H1299 cells (C) and in indicated A549 cells
   (D) treated with 50 nM nab-PTX for 16 h. Green: β-tubulin; blue: nuclei
   with DAPI. E. Cell cycle measurement in indicated H1299 cells and in
   indicated A549 cells treated with 50 nM nab-PTX for 16 h. F-G.
   Statistical analyses of cell cycle measurements in indicated
   H1299 cells (F) and in indicated A549 cells (G) treated as E. sgC,
   sgCtrl; sgL, sgLKB1; EV, empty vector; WT, wide type; KD, kinase dead.
   Scale bar, 10 μM. ns, no statistical significance. The mean ± s.d. are
   shown; n = 3 independent experiments. Statistical analysis was
   performed using an unpaired, two-tailed t-test.

3.3. AMPK inactivation sensitizes cells to apoptotic cell death induced by
nab-PTX in LKB1-WT NSCLC

   To determine the mechanisms by which LKB1 impairs the efficacy of
   nab-PTX, RNA sequencing was performed to examine the changes in gene
   expression profiles after genetic alteration of LKB1 in H1299 and
   A549 cells. By employing differentially expressed genes (DEGs)
   analysis, we observed that the overexpression of WT-LKB1 significantly
   upregulated 397 genes and downregulated 226 genes in A549 cells
   ([124]Fig. S2A). Subsequently, we executed Hallmark-pathway enrichment
   analysis, and the top 10 gene sets were screened out, including
   epithelial mesenchymal transition, TGFβ signaling and hypoxia
   signialing (EMT), ([125]Fig. S2B). We found that these screened
   pathways mostly require AMPK, which is a primary downstream substrate
   of LKB1 and plays a critical role in metabolism and cell death
   [[126]24,[127]39,[128]40]. Therefore, we conducted gene set enrichment
   analysis (GSEA) and found that the overexpression of LKB1 in A549 cells
   indeed significantly activated AMPK signaling ([129]Fig. 3A). In
   contrast, the deletion of LKB1 impaired AMPK signaling in H1299 cells
   ([130]Fig. S2C). Similar observations were obtained in H1650 cells,
   whose gene expression profiles ([131]GSE69747) were acquired from GEO
   database ([132]Fig. S2D). To further verify the results from the
   bioinformatics analyses, we directly measured the level of AMPK
   activation in a panel of NSCLC cell lines by immunoblotting.
   Consistently, higher levels of phosphorylation of AMPK in the T172
   residue (p-AMPK) and ACC in the S79 residue (p-ACC) were found in the
   NSCLC cell lines with WT-LKB1 in comparison to the cell lines with
   mutant LKB1 ([133]Fig. 3B). Furthermore, the deletion of LKB1
   strikingly reduced p-AMPK and p-ACC in H1299 cells ([134]Fig. 3C),
   whereas the overexpression of WT-LKB1, but not the KD form,
   dramatically upregulated p-AMPK and p-ACC in A549 cells ([135]Fig. 3D).

Fig. 3.

   [136]Fig. 3
   [137]Open in a new tab

   AMPK inactivation sensitizes cells to apoptotic cell death induced by
   nab-PTX.

   A. Gene set enrichment analysis (GSEA) using the Pathway Interaction
   Databases (PID) and Wikipathways database focusing on LKB1 signaling
   and AMPK signaling gene sets, respectively. FDR, false discovery rate;
   NES, normalized enrichment score. B. An immunoblot showing the levels
   of AMPK T172 phosphorylation, AMPK, ACC S79 phosphorylation and ACC in
   various NSCLC cell lines. C-D. An immunoblot showing the levels of AMPK
   T172 phosphorylation, AMPK, ACC S79 phosphorylation and ACC in
   indicated H1299 cells (C) and in indicates A549 cells (D). E. An
   immunoblot showing the levels of AMPK T172 phosphorylation, AMPK, ACC
   S79 phosphorylation and ACC in H1299 cells treated with 50 nM nab-PTX
   combined with or without 10 μM BAY-3827 for 24 h. F. Cell viability
   measurement in H1299 cells treated with indicated concentrations of
   nab-PTX combined with or without 10 μM BAY-3827 for 72 h. G. Cell death
   measurement in H1299 cells treated with 50 nM nab-PTX combined with or
   without 10 μM BAY-3827 for 36 h. H. An immunoblot indicating the loss
   of AMPKα1/α2 in AMPK DKO H1299 cell lines generated using CRISPR–Cas9
   system. I. Cell viability measurement in AMPK WT and DKO H1299 cells
   treated with indicated concentrations of nab-PTX for 72 h. J. Cell
   death measurement in AMPK WT and DKO H1299 cells treated with 50 nM
   nab-PTX for 36 h. K. Cell death measurement in AMPK WT and DKO
   H1299 cells treated with 50 nM nab-PTX with or without 40 μM EX229 for
   36 h. L. Immunoblot showing the levels of AMPK T172 phosphorylation,
   AMPK, ACC S79 phosphorylation, ACC, cleaved-caspase 7 (C-Cas-7), and
   cleaved PARP (C-PARP) in AMPK WT and DKO H1299 cells treated as in K.
   M. Cell death measurement in AMPK DKO H1299 cells treated with 50 nM
   nab-PTX, 40 μM EX229, and various cell death inhibitors. Z-V, 20 μM
   Z-VAD-FMK; Fer-1, 5 μM ferrostatin-1; GSK, 20 μM GSK-872. sgC, sgCtrl;
   EV, empty vector; WT, wild-type. ns, not statistically significant;
   ∗∗∗∗p < 0.0001. The mean ± s.d. is shown as n ≥ 3 independent
   experiments. Statistical analyses were performed using unpaired
   two-tailed t-tests.

   To explore whether AMPK is involved in the nab-PTX resistance induced
   by LKB1 in NSCLC, we firstly treated the cells with nab-PTX combined
   with or without BAY-3827, a specific and potent AMPK inhibitor
   [[138]41,[139]42]. Reduced levels of p-ACC upon treatment with BAY-3827
   indicated the efficient inhibition of AMPK ([140]Fig. 3E). Notably, the
   pharmacological inhibition of AMPK by BAY-3827 significantly enhanced
   the cytotoxicity of nab-PTX in H1299 ([141]Fig. 3F and G) and
   H358 cells ([142]Figs. S2E and S2F). Moreover, using CRISPR/Cas9
   technology, we established an AMPKα1/AMPKα2 double knockout (AMPK-DKO)
   in H1299 cells and found that the deletion of AMPK dramatically
   sensitized the cells to the treatment of nab-PTX ([143]Fig. 3H–J). To
   further verify the role of AMPK, we tested the effect of EX229, an
   allosteric AMPK activator, and found that the addition of EX229
   significantly protected cells from cell death induced by nab-PTX only
   in AMPK-WT cells, but not in AMPK-DKO cells ([144]Fig. 3K and L).
   Z-VAD-FMK, a general caspase inhibitor, significantly reduced the cell
   death induced by nab-PTX in AMPK-DKO cells ([145]Fig. 3M). Notably, a
   relatively lower p-AMPK level in H1975 cells could be observed, which
   is likely the cause for its sensitivity to nab-PTX in comparison to
   other LKB1-WT NSCLC cell lines ([146]Fig. 1, [147]Fig. 3B). Taken
   together, these results indicate that AMPK activation downstream of
   LKB1 exerts protective effect, and AMPK inactivation sensitizes cells
   to apoptotic cell death induced by nab-PTX.

3.4. LKB1 impairs nab-PTX-induced apoptotic cell death through AMPK
activation

   To further verify whether AMPK activation plays a causal role in
   determining the sensitivity of nab-PTX downstream of LKB1, we tested a
   series of AMPK activators, including AICAR, metformin, A-769662 and
   EX229. Based on the mechanism by which they activate AMPK, these
   activators can be divided into two groups: AICAR and metformin are
   LKB1-dependent and regulated by the intracellular AMP/ATP ratio, while
   A-769662 and EX229 are LKB1-independent and act allosterically by
   binding to the β-subunit of AMPK [[148]43]. In H1299 cells with
   WT-LKB1, both LKB-dependent and LKB1-independent AMPK activators
   significantly enhanced the activation of AMPK ([149]Fig. 4A) and
   reduced the cell death induced by nab-PTX ([150]Fig. 4B). As expected,
   in A549 cells with mutated LKB1, only LKB1-independent AMPK activators
   (A-769662 and EX229) activated AMPK ([151]Fig. 4C) and displayed a
   protective effect ([152]Fig. 4D). Moreover, we reconstituted
   constitutively activated (CA) AMPK (amino acids 1–312 of AMPKα1) in
   LKB1 KO H1299 cells and found that overexpression of CA-AMPK moderately
   increased the phosphorylation of ACC ([153]Fig. 4E) and partially
   reduced the cell death induced by nab-PTX ([154]Fig. 4F). A similar
   function of CA-AMPK was observed in A549 cells in which LKB1 is mutant
   ([155]Fig. 4G and H). Moreover, the deletion of LKB1 in H1299 cells did
   not alter the effects of EX229, a LKB1-independent AMPK activator, on
   AMPK activation and apoptotic cell death induced by nab-PTX ([156]Fig.
   4I and J). Similarly, the protective effect of EX229 against
   nab-PTX-induced cell death was observed in A549 cells regardless of the
   status of LKB1 ([157]Fig. 4K and L). Above all, these data indicate
   that LKB1 inhibits nab-PTX-induced apoptotic cell death through AMPK
   activation.

Fig. 4.

   [158]Fig. 4
   [159]Open in a new tab

   LKB1 inhibits nab-PTX induced apoptotic cell death through AMPK
   activation.

   A. An immunoblot showing the levels of AMPK T172 phosphorylation, AMPK,
   ACC S79 phosphorylation, ACC in H1299 cells treated with 200 μM
   A-769662, 40 μM EX229, 200 μM AICAR and 1 mM Metformin for 24 h,
   respectively. B. Cell death measurement in H1299 cells treated with
   50 nM nab-PTX combined with or without indicated AMPK activators for
   36 h. C. An immunoblot showing the levels of AMPK T172 phosphorylation,
   AMPK, ACC S79 phosphorylation, ACC in A549 cells treated as in A. D.
   Cell death measurement in A549 cells treated with 50 nM nab-PTX
   combined with or without indicated AMPK activators for 36 h. E. An
   immunoblot showing the expression of CA-AMPK and ACC S79
   phosphorylation in LKB1 KO H1299 cells. F. Cell death measurement in
   indicated H1299 cells upon the treatment of 50 nM nab-PTX for 36 h. G.
   An immunoblot showing the expression of CA-AMPK and ACC S79
   phosphorylation in A549 cells. H. Cell death measurement in indicated
   A549 cells treated with 50 nM nab-PTX for 36 h. I. An immunoblot
   showing AMPK T172 phosphorylation, AMPK, ACC S79 phosphorylation, ACC,
   C-Cas-7, and C-PARP in indicated H1299 cells treated with 50 nM nab-PTX
   combined with or without 40 μM EX229 for 24 h. J. Cell death
   measurement in indicated H1299 cells treated with 50 nM nab-PTX
   combined with or without 40 μM EX229 for 36 h. K. An immunoblot showing
   AMPK T172 phosphorylation, AMPK, ACC S79 phosphorylation, ACC, C-Cas-7,
   and C-PARP in indicated A549 cells treated as in I. L. Cell death
   measurement in indicated A549 cells treated as in J. C-Cas-7,
   cleaved-caspase 7; C-PARP, cleaved-PARP. sgC, sgCtrl; EV, empty vector;
   WT, wide type; KD, kinase dead. ns, no statistical significance;
   ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001. The mean ± s.d.
   are shown; n ≥ 3 independent experiments. Statistical analysis was
   performed using an unpaired, two-tailed t-test.

3.5. Activation of the LKB1-AMPK signaling pathway inhibits nab-PTX-induced
apoptosis by impairing ROS production

   We next sought to study the underlying mechanism(s) by which the
   LKB1-AMPK signaling pathway leads to the resistance to nab-PTX. Given
   that AMPK can reduce reactive oxygen species (ROS) under several stress
   conditions [[160][44], [161][45], [162][46]] and ROS is one of the key
   effectors of the anticancer efficacy of PTX [[163]47,[164]48], we
   deduced that the activation of AMPK can reduce the ROS level induced by
   nab-PTX, which partially protects the cells from the apoptotic cell
   death. To test this hypothesis, we first detected the ROS level in
   cells treated with nab-PTX. As showed in [165]Fig. 5A–B, nab-PTX
   significantly increased the ROS level in A549 cells. To confirm the
   proapoptotic role of ROS, N-Acetylcysteine (NAC), a well-known and
   effective ROS scavenger [[166]49], was applied for the experiment. As
   expected, NAC significantly reduced the ROS level ([167]Fig. 5C and D)
   and repressed the apoptotic cell death ([168]Fig. 5E and F) induced by
   nab-PTX in A549 cells. Moreover, the allosteric activator of AMPK,
   EX229, significantly reduced the ROS level induced by nab-PTX in
   A549 cells ([169]Fig. 5G and H). In addition, overexpression of WT-LKB1
   in A549 cells led to lower level of ROS production induced by nab-PTX
   ([170]Fig. 5I and J), and consistently NAC displayed a weaker
   inhibitory effect on ROS level in LKB1-WT A549 cells ([171]Fig. 5I and
   J). Correspondingly, NAC exhibited a relatively weaker protective
   effect against cell death induced by nab-PTX in LKB1-WT A549 cells.
   ([172]Fig. 5K and L). Conversely, BAY-3827, an AMPK inhibitor,
   significantly increased the ROS level ([173]Fig. 5M and N) and promoted
   the apoptotic cell death induced by nab-PTX ([174]Fig. 5O and P) in
   LKB1-WT A549 cells. Taken together, our data suggests that the
   activation of LKB1-AMPK signaling pathway inhibits apoptosis induced by
   nab-PTX through the regulation of ROS production and oxidative stress.

Fig. 5.

   [175]Fig. 5
   [176]Open in a new tab

   Activation of the LKB1-AMPK signaling pathway inhibits nab-PTX induced
   apoptosis by impairing ROS production.

   A. Representative images showing the production of ROS in A549 cells
   treated with 50 nM nab-PTX for indicated time point. B. ROS measurement
   in A549 cells treated as in A. C. Representative images showing the
   productions of ROS in A549 cells treated with 50 nM nab-PTX combined
   with or without 10 mM NAC for 4 h. D. ROS measurement in A549 cells
   treated as in C. E. Cell death measurement in A549 cells treated with
   50 nM nab-PTX combined with or without 10 mM NAC for 36 h. F. An
   immunoblot showing the expression of C-Cas-7 and C-PARP in A549 cells
   treated with 50 nM nab-PTX combined with or without 10 mM NAC for 24 h.
   G. Representative images showing the productions of ROS in A549 cells
   treated with 50 nM nab-PTX combined with or without 40 μM EX229 for
   4 h. H. ROS measurement in A549 cells treated as in G. I.
   Representative images showing the productions of ROS in indicated
   A549 cells treated with 50 nM nab-PTX combined with or without 10 mM
   NAC for 4 h. J. ROS measurement in indicated A549 cells treated as in
   I. K. Cell death measurement in indicated A549 cells treated with 50 nM
   nab-PTX combined with or without 10 mM NAC for 36 h. L. An immunoblot
   showing the expression of C-Cas-7 and C-PARP in indicated A549 cells
   treated with 50 nM nab-PTX combined with or without 10 mM NAC for 24 h.
   M. Representative images showing the productions of ROS in indicated
   A549 cells treated with 50 nM nab-PTX combined with or without 10 μM
   BAY-3827 for 4 h. N. ROS measurement in indicated A549 cells treated as
   in M. O. Cell death measurement in indicated A549 cells treated with
   50 nM nab-PTX combined with or without 10 μM BAY-3827 for 36 h. P. An
   immunoblot showing the expression of C-Cas-7 and C-PARP in indicated
   A549 cells treated with 50 nM nab-PTX combined with or without 10 μM
   BAY-3827 for 24 h. C-Cas-7, cleaved-caspase 7; C-PARP, cleaved-PARP;
   EV, empty vector; WT, wide type. ns, no statistical significance;
   ∗p < 0.05; ∗∗∗∗p < 0.0001. The mean ± s.d. are shown; n ≥ 3 independent
   experiments. Statistical analysis was performed using an unpaired,
   two-tailed t-test.

3.6. Upregulation of the LKB1-AMPK-SLC7A11-GSH pathway mediates nab-PTX
resistance in LKB1-WT NSCLC cells

   To further explore the mechanism by which LKB1-AMPK activation inhibits
   the production of ROS and cell death induced by nab-PTX, we first
   performed gene set enrichment analysis and found that the
   overexpression of WT LKB1 significantly enhanced the antioxidant
   signaling pathway in A549 cells ([177]Fig. S3A). In contrast, the
   deletion of LKB1 obviously impaired the antioxidant signaling in
   H1299 cells ([178]Fig. S3B). Given that glutathione (GSH) is the major
   antioxidant to prevent cells from the damage of ROS [[179]50], we next
   measured the level of intracellular GSH after the genetic alteration of
   LKB1 in H1299 and A549 cells. Results showed that the deletion of LKB1
   significantly decreased the GSH level in H1299 cells ([180]Fig. 6A),
   while the overexpression of WT LKB1 but not its kinase dead form
   increased the GSH level in A549 cells ([181]Fig. 6B). To verify the
   role of GSH in the cell death induced by nab-PTX, we tested the effect
   of dl-Buthionine-(S,R)-sulfoximine hydrochloride (BSO), a potent
   inhibitor of γ-glutamylcysteine synthetase (GCS, the first enzyme of
   the cellular GSH biosynthetic pathway). As expected, a significant
   reduction of cellular GSH level was observed upon the treatment of BSO
   ([182]Fig. 6C). Meanwhile, BSO strikingly enhanced cell death caused by
   nab-PTX in H1299 cells ([183]Fig. 6D and E), demonstrating the
   important role of GSH in protecting the apoptotic cell death induced by
   nab-PTX.

Fig. 6.

   [184]Fig. 6
   [185]Open in a new tab

   Upregulation of the LKB1-AMPK-SLC7A11-GSH pathway mediates nab-PTX
   resistance in LKB1-WT NSCLC cells.

   A-B. Cellular GSH measurement in indicated H1299 (A) and A549 (B)
   cells. C. Cellular GSH measurement in H1299 cells treated with
   indicated concentration of BSO for 24 h. D. Cell death measurement in
   H1299 cells treated with 50 nM nab-PTX in combination with or without
   50 μM BSO for 36 h. E. An immunoblot showing the expression of C-Cas-7
   and C-PARP in H1299 cells treated with 50 nM nab-PTX combined with or
   without 50 μM BSO for 24 h. F. Graphical summary of GSH metabolism.
   G-H. mRNA levels of GSH metabolism-related genes in indicated H1299 (G)
   and A549 (H) cells. I-J. An immunoblot showing the expression of
   SLC7A11 in indicated H1299 (I) and A549 (J) cells. K. An immunoblot
   showing the expression of SLC7A11 in indicated H1299 cells. L. Cellular
   GSH measurement in indicated H1299 cells. M. An immunoblot showing the
   ACC S79 phosphorylation and the expression of SLC7A11 in indicated
   H1299 cells. N. Cellular GSH measurement in indicated H1299 cells. O.
   An immunoblot showing the ACC S79 phosphorylation and the expression of
   SLC7A11 in indicated H1299 cells treated with or without 40 μM EX229
   for 24 h. P. Cellular GSH measurement in indicated H1299 cells treated
   with or without 40 μM EX229 for 24 h. Q. An immunoblot showing the
   downregulation of SLC7A11 in SLC7A11-knockdown H1299 cells. R. Cellular
   GSH measurement in indicated H1299 cells. S-T. ROS measurement (S) and
   cell death measurement (T) in indicated H1299 cells treated with 50 nM
   nab-PTX for 4 h or 36 h, respectively. U. An immunoblot showing the
   expression of C-Cas-7 and C-PARP in indicated H1299 cells treated with
   50 nM nab-PTX for 24 h. V. An immunoblot showing the expression of
   SLC7A11 in indicated H1299 cells. W. Cellular GSH measurement in
   indicated H1299 cells. X–Y. ROS measurement (X) and cell death
   measurement (Y) in indicated H1299 cells treated with 50 nM nab-PTX for
   4 h or 36 h, respectively. Z. An immunoblot showing the expression of
   C-Cas-7 and C-PARP in indicated H1299 cells treated with 50 nM nab-PTX
   for 24 h. C-Cas-7, cleaved-caspase 7; C-PARP, cleaved-PARP. EV, empty
   vector; WT, wide type. ns, no statistical significance; ∗p < 0.05;
   ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001. The mean ± s.d. are shown;
   n ≥ 3 independent experiments. Statistical analysis was performed using
   an unpaired, two-tailed t-test.

   We then tried to investigate the molecular mechanism by which LKB1
   increases the cellular GSH level. It has been well established the
   intracellular GSH level is tightly regulated by a group of enzymes:
   SLC7A11, the transporter of cystine which is a substrate of GSH
   biosynthesis; γ-glutamylcysteine synthetase (GCS) and glutathione
   synthetase (GSS), who are responsible for the De novo synthesis of GSH;
   Glutathione peroxidase 4 (GPX4), which catalyzes GSH to glutathione
   disulfide (GSSG) and glutathione reductase (GR), which restores GSH by
   catalyzing the reverse reaction [[186]51]([187]Fig. 6F). In this study,
   we found that the deletion of LKB1 significantly downregulated the mRNA
   level of SLC7A11 in H1299 cells ([188]Fig. 6G) whereas overexpression
   of LKB1 increases its mRNA level in A549 cells ([189]Fig. 6H).
   Furthermore, similar changes of SLC7A11 protein levels could also be
   observed in H1299 and A549 cells ([190]Fig. 6I and J). Moreover, we
   found an obvious downregulation of SLC7A11 in AMPK-DKO cells compared
   to the AMPK-WT cells ([191]Fig. 6K), suggesting the causal role of AMPK
   in regulation of SLC7A11 downstream of LKB1. Consistently, double
   knockout of AMPK significantly reduced the cellular GSH level
   ([192]Fig. 6L). In addition, the overexpression of a constitutively
   active form of AMPK (CA-AMPK) apparently restored the expression of
   SLC7A11 and increased the cellular GSH level in LKB1 KO H1299 cells
   ([193]Fig. 6M and N). In addition to the genetic approaches, we also
   used a pharmacological reagent, EX229, an LKB1-independent AMPK
   activator, and found that EX229 dramatically upregulated the expression
   of SLC7A11 and increased the cellular GSH level both in LKB1 WT and
   LKB1 KO H1299 cells ([194]Fig. 6O and P). Interestingly, only the
   LKB1-independent AMPK activators (A-769662 and EX229) but not the
   LKB1-dependent AMPK activators displayed a potent effect on the
   upregulation of SLC7A11 in A549 cells ([195]Fig. S3C), further
   demonstrating the critical role of LKB1-AMPK signaling in determining
   the cell death induced by nab-PTX.

   Based on the above results, we reasoned that SLC7A11 might play a
   positive role in the primary resistance of nab-PTX in LKB1-WT NSCLC. To
   verify our hypothesis, we knocked down (KD) SLC7A11 in H1299 cells
   ([196]Fig. 6Q) and found dramatic reduction of cellular GST level
   ([197]Fig. 6R) in cells with KD of SLC7A11. As expected, a higher ROS
   level and more cell death could be observed in SLC7A11 KD H1299 cells
   after the treatment of nab-PTX ([198]Fig. 6S–U). Meanwhile, we also
   overexpressed SLC7A11 in LKB1 KO H1299 cells ([199]Fig. 6V). The
   overexpression of SLC7A11 significantly restored the cellular GST level
   ([200]Fig. 6W), reduced the ROS accumulation ([201]Fig. 6X), and
   impaired the cell death induced by nab-PTX ([202]Fig. 6Y and Z).
   Similar effects of SLC7A11-overexpression could also be observed in
   A549 cells ([203]Figs. S3D–F).

   Taken together, data from this part of our study suggest that the
   activation of LKB1-AMPK signaling pathway increases cellular GSH level
   by promoting SLC7A11 expression, which attenuates the efficacy of
   nab-PTX through the regulation of ROS production and cell death.

3.7. Inhibition of AMPK sensitizes LKB1-WT NSCLC to nab-PTX in vivo

   To evaluate whether the inhibition of AMPK can promote the therapeutic
   potency of nab-PTX in vivo, we established nude mice xenografts using
   H1299 and H358 cells, both with WT-LKB1, in which the animals were
   treated with nab-PTX with or without the AMPK inhibitor BAY-3827. To
   visualize tumor growth in vivo, lentiviruses expressing EGFP were
   infected to these cells before subcutaneous injection into the mice.
   Combined treatment with nab-PTX and BAY-3827 displayed a stronger
   therapeutic effect as measured by tumor growth and tumor volume in
   H1299-derived xenografts ([204]Fig. 7A–C). Similarly, the combination
   treatment of nab-PTX and BAY-3827 also showed a better therapeutic
   effect in H358-derived xenografts than nab-PTX alone ([205]Fig. 7D–F).
   The combined effect of nab-PTX and BAY-3827 was recapitulated by
   immunohistochemistry (IHC) staining analysis of the xenograft tumors,
   in which p-ACC was abrogated by BAY-3827 in both H1299- and
   H358-derived xenografts ([206]Fig. 7G -J). BAY-3827 also enhanced the
   efficacy of nab-PTX in inhibiting tumor cell proliferation and
   increasing tumor cell death, as quantified by Ki-67 and p-H2A.X
   staining, respectively ([207]Fig. 7G-J). Taken together, these data
   suggest that AMPK inhibition sensitizes LKB1-WT NSCLC to nab-PTX.

Fig. 7.

   [208]Fig. 7
   [209]Open in a new tab

   Inhibition of AMPK increases therapeutic potential of nab-PTX in
   LKB1-WT NSCLC.

   A. Representative images showing the tumor burden assessed by
   bioimaging in mice xenografted with H1299 cells upon the treatments of
   10 mg/kg nab-PTX with or without 10 mg/kg BAY-3827. B-C. Tumor volume
   measurements over the course of the treatment (B) and final tumor
   weight measurement (C) in H1299-derived xenografts upon the treatment
   as in A. D. Representative images showing the tumor burden assessed by
   bioimaging in mice xenografted with H358 cells upon the treatments of
   10 mg/kg nab-PTX with or without 10 mg/kg BAY-3827. E-F. Tumor volume
   measurements over the course of the treatment (E) and final tumor
   weight measurement (F) in H358-derived xenografts upon the treatment as
   in D. G-H. Representative immunohistological images showing the
   expression of ACC S79 phosphorylation, Ki-67, H2A.X S139
   phosphorylation (G) and correlated quantifications (H) in H1299-derived
   xenografts upon the treatment as in A. I-J. Representative
   immunohistology images showing the expression of ACC S79
   phosphorylation, Ki-67, H2A.X S139 phosphorylation (I) and correlated
   quantifications (J) in H358-derived xenografts upon the treatment as in
   D. Combo, nab-PTX combined with BAY-3827. Scale bar, 100 μM. ns, no
   statistical significance; ∗p < 0.05; ∗∗p < 0.01; ∗∗∗∗p < 0.0001. The
   mean ± s.d. are shown; n ≥ 3 independent experiments. Statistical
   analysis was performed using an unpaired, two-tailed t-test.

4. Discussion

   Chemotherapy is usually considered as a non-targeted treatment, as
   chemotherapeutic agents usually act via direct inhibition of DNA
   synthesis or causing damage to DNA and proteins [[210]52].
   Conceptually, one chemotherapeutic drug should be available for the
   treatment of all kinds of cancer. However, the application of a
   chemotherapeutic drug is usually limited to not more than 5 types of
   cancer. For instance, nab-PTX has been approved by the FDA for the
   treatment of metastatic breast cancer, locally advanced or metastatic
   NSCLC and advanced pancreatic cancer [[211]16]. Similarly, cisplatin,
   another well-known chemotherapeutic drug, is approved for the treatment
   of advanced ovarian cancer, testicular cancer, and bladder cancer
   [[212]53,[213]54]. Therefore, it remains a challenge whether a
   chemotherapeutic drug such as nab-PTX can also be targeted to a
   specific group of cancer patients with a specific genetic background.
   In this study, we found that NSCLC with mutant-LKB1 is more sensitive
   to nab-PTX in comparison to those NSCLC with WT-LKB1. Mechanistically,
   LKB1 status does not alter the intracellular concentration of nab-PTX
   or affect its canonical pharmacological action in promoting microtubule
   polymerization. The deficiency of AMPK activation reduces the
   expression of SLC7A11 and leads to a lower cellular GSH level, which
   sensitizes LKB1-mutant NSCLC to oxidative stress and cell death induced
   by nab-PTX. On the other hand, genetic or pharmacological inhibition of
   AMPK is capable of sensitizing LKB1-WT NSCLC to nab-PTX ([214]Fig. 8).
   Therefore, our data reveal a novel role of LKB1-AMPK-SLC7A11-GSH
   signaling pathway in the primary resistance to nab-PTX, and provide a
   therapeutic strategy for the treatment of LKB1-WT NSCLC by targeting
   the LKB1-AMPK-SLC7A11 pathway.

Fig. 8.

   [215]Fig. 8
   [216]Open in a new tab

   Schematic model showing the mechanism by which LKB1 attenuates the
   efficacy of nab-PTX.

   Activation of the LKB1-AMPK signaling pathway increases cellular GSH
   level by promoting the expression of SLC7A11, which impairs the
   therapeutic potential of nab-PTX for the reduction of ROS in LKB1-WT
   NSCLC cells. The combination treatment of AMPK inhibitor, BAY-3827,
   efficiently sensitized NSCLC cells to nab-PTX treatment both in vitro
   and in vivo.

   An earlier report by Li R et al. found that nab-PTX has higher
   selectivity towards murine Kras-mutant NSCLC via increased uptake of
   nab-PTX, a process associated with enhanced level of AMPK activation
   [[217]33]. As LKB1 is the key upstream kinase of AMPK, we attempted to
   explore whether the LKB1-AMPK pathway would also affect the cellular
   uptake of nab-PTX in NSCLC. Intriguingly, we found that the LKB1-AMPK
   status has no evident effect on the cellular uptake of nab-PTX, neither
   alters its effects on microtubule and cell cycle. We believe that this
   discrepancy may be caused by different contexts. Actually, a recent
   study has pointed out the distinct role of LKB1 in the development of
   human and murine NSCLC: murine NSCLC cells with mutated LKB1 exhibited
   an advantage in tumor growth in vivo, whereas mutation of LKB1 in human
   NSCLC cells impaired tumor growth in a mouse xenograft model [[218]55].

   In addition to AMPK, other 12 AMPK-related kinase (MARK1, MARK2, MARK3,
   MARK4, SAD-A, SAD-B, QIK, QSK, NUAK1, NUAK2, SIK, and SNRK) that
   contain the LXT motif in their activation loop have also been
   identified as the substrates of LKB1 [[219]56]. Among them, microtubule
   affinity regulating kinase (MARK) 1/2/3/4 and synapses of amphids
   defective (SAD)-A/B have been reported to regulate tubulin
   polymerization through phosphorylation of microtubule-associated
   proteins (MAPs), such as Tau, MAP2, and MAP4 [[220]57]. Therefore, LKB1
   may play a regulatory role in microtubule dynamics. Recently, Kojima et
   al. reported that LKB1 can directly phosphorylate and activate MARK2,
   which in turn phosphorylates the microtubule-associated protein Tau and
   suppresses tubulin polymerization [[221]58]. Meanwhile, Mian et al.
   also reported that the overexpression of LKB1 destabilizes microtubule
   in myoblast cells [[222]59]. However, in our study, the LKB1 status
   does not show a dramatic effect on the changes of microtubule dynamics
   induced by nab-PTX, thus excluding the possibility that the high
   susceptibility of LKB1-mutant NSCLC cells to nab-PTX is caused by
   alterations in microtubule polymerization induced by nab-PTX in NSCLC
   cells.

   At present, the relationship between ROS/oxidative stress and AMPK
   activation is intricate. From the perspective of the impact of ROS on
   AMPK, it seems that ROS play a dual role in AMPK activation. On one
   hand, ROS are usually considered as a trigger of AMPK signaling,
   because high level of ROS impairs mitochondrial function and increases
   the cellular AMP level, which subsequently promotes AMPK activation
   [[223]60,[224]61]. Meanwhile, it is also reported that the elevation of
   ROS can directly oxidize redox-sensitive cysteine residues on the α
   subunit of AMPK, leading to AMPK activation [[225]62]. On the other
   hand, some studies demonstrated that ROS and oxidative stress impair
   AMPK activation via direct oxidation of its key residues or protein
   degradation [[226]63,[227]64]. It appears that the different effects of
   ROS/oxidative stress on AMPK activation may be due to the distinction
   in cellular contents, the amount of ROS and the types of ROS.
   Similarly, AMPK also plays a dual role in ROS/oxidative stress.
   Generally, the activation of AMPK is considered to compromise
   ROS/oxidative stress under different stress conditions
   [[228]25,[229]31,[230]65,[231]66]. However, some studies also reported
   that AMPK activity contributes to oxidative stress. For instance, Song
   et al., demonstrated that AMPK-mediated phosphorylation of BECN1 at
   Ser90/93/96 is necessary for the formation of BECN1-SLC7A11 complex,
   which promotes the lipid peroxidation upon the treatment of SLC7A11
   inhibitors [[232]67].

   In this study, one important finding is that activated AMPK enhances
   the expression of SLC7A11. Recently, some studies reported that
   activated AMPK promotes SLC7A11 expression via phosphorylation and
   stabilization of NRF2 that is considered as a transcription factor of
   SLC7A11 [[233]68,[234]69]. Meanwhile, Wang et al. suggested that AMPK
   increases the protein stability of SLC7A11 by promoting its
   palmitoylation-mediated by ZDHHC8 [[235]70]. More work is needed to
   establish the molecular mechanisms linking AMPK activation to SLC7A11
   expression.

   Based on the critical role of AMPK in various cellular processes
   including metabolism and cell death, significant progresses have been
   made in development of AMPK activators for treatment of obesity and
   diabetes [[236]71]. In contrast, studies on developing AMPK inhibitors
   are stagnant. For a long period of time, Compound-C has been the only
   pharmacological inhibitor of AMPK for scientific research [[237]41].
   Recently, a novel and potent AMPK inhibitor, BAY-3827, has been
   developed [[238]42], which displays a more than 100-fold inhibitory
   potency and a better selectivity to AMPK, in comparison to Compound-C
   [[239]41]. In our study, the inhibitory effect of BAY-3827 on AMPK
   activity could be proved by the almost complete abolition of ACC S79
   phosphorylation ([240]Fig. 3E, [241]Fig. S2E). Meanwhile, also it is
   intriguing to observe that there was a dramatic increase in AMPK T172
   phosphorylation upon the treatment of BAY-3827 both in H1299 and
   H358 cells ([242]Fig. 3E, [243]Fig. S2E). A recent study also reported
   similar results induced by BAY-3827, due to the inhibitory effect of
   BAY-3827 on AMPK T172 dephosphorylation [[244]41].

   In summary, our study reveals a novel function of the
   LKB1-AMPK-SLC7A11-GSH signaling pathway in the regulation of primary
   resistance to nab-PTX in NSCLC and provides a rationale for the
   combination of nab-PTX with AMPK inhibitor in LKB1-WT NSCLC to achieve
   better therapeutic efficacy.

CRediT authorship contribution statement

   Dade Rong: Data curation, Formal analysis, Investigation, Methodology,
   Project administration, Validation, Writing – original draft, Writing –
   review & editing. Liangliang Gao: Data curation, Investigation,
   Methodology. Yiguan Chen: Data curation, Investigation, Methodology.
   Xiang-Zheng Gao: Data curation, Investigation, Methodology. Mingzhu
   Tang: Investigation, Methodology.Haimei Tang: Investigation,
   Methodology. Yuan Gao: Investigation, Methodology. Guang Lu: Formal
   analysis, Writing – review & editing. Zhi-Qiang Ling: Methodology,
   Resources, Writing – review & editing. Han-Ming Shen: Methodology,
   Resources, Writing – review & editing.

Data access statement

   All relevant data are within the paper and its Supporting Information
   files.

Funding

   This project was supported by the following research grants to Shen HM
   from the University of Macau (CPG2023-0032-FHS, CPG2024-0035-FHS
   UM-MYRG2020-00022-FHS) and the Macau Science and Technology Development
   Fund (FDCT0078/2020/A2, 0031/2021/A1, FDCT0081/2022/AMJ, FDCT
   0004/2021/AKP); National Key R&D Program of China (No. 2022YFE0205900)
   to Ling ZQ.

Declaration of competing interest

   The authors declare no competing interests.

Acknowledgements