Abstract
Overcoming acquired resistance to Osimertinib remains a critical
challenge in treating NSCLC. This research indicates that
Osimertinib‐resistant cells exhibit a strong dependence on glutamine
metabolism. However, targeting GLS1 shows limited anticancer effects,
probably because it cannot fully block the glutamine metabolic pathway.
The investigation reveals that a more effective strategy involves
simultaneously inhibiting both ASCT2 and GLS1. After confirming the
efficacy of this dual‐targeting approach against Osimertinib‐resistant
cells in preclinical models, the potential of utilizing a
broad‐spectrum glutamine metabolism antagonist is further explored to
achieve superior antitumor efficacy. DON, broad‐spectrum glutamine
antagonist, presents toxicity issues. Herein, the high NQO1 expression
in Osimertinib‐resistant NSCLC cells is leveraged to design an
NQO1‐responsive DON prodrug, 10e (LBJ‐10e). This prodrug demonstrates
superior safety compared to natural DON and greater antitumor activity
against resistant tumors compared to the clinical phase II drug DRP104.
These findings may address the clinical limitations of GLS1 allosteric
inhibitors and underscore prodrug strategies in effectively treating
Osimertinib‐resistant lung cancer, providing a foundation for future
clinical trials.
Keywords: drug design, glutamine, NQO1, NSCLC, Osimertinib resistance
__________________________________________________________________
Glutamine metabolism in Osimertinib‐resistant NSCLC is complex and
multifaceted; therefore, a comprehensive inhibition strategy targeting
multiple nodes, such as ASCT2 and GLS1, or employing broad‐spectrum
inhibition is essential to prevent rapid metabolic rewiring. The
NQO1‐activated DON prodrug shows significant promise in this context,
underscoring its potential to achieve a favorable therapeutic index.
graphic file with name ADVS-12-2411479-g002.jpg
1. Introduction
The most recent research has uncovered that the mortality trends of
cancer are predominantly shaped by lung cancer, which consistently
accounts for significantly higher annual deaths compared to the
combined mortality of colorectal, breast, and prostate cancer.^[ [44]^1
^] Non‐small cell lung cancer (NSCLC), the most prevailing histological
subtype of lung cancer, comprises ≈85% of all patients.^[ [45]^2 ^]
Typically, surgical resection is the standard therapeutic approach for
early‐stage NSCLC, while patients with advanced NSCLC predominantly
receive pharmacological therapy. Epidermal growth factor receptor
tyrosine kinase inhibitor (EGFR‐TKI) represents the primary choice for
patients with advanced NSCLC harboring EGFR mutations. Osimertinib, an
efficient and irreversible third‐generation EGFR‐TKI, is prescribed as
a standard first‐line treatment for patients with advanced or
metastatic NSCLC bearing EGFR mutations due to its remarkable antitumor
efficacy and manageable safety profile.^[ [46]^3 , [47]^4 , [48]^5 ,
[49]^6 ^] However, similar to other EGFR‐TKIs, Osimertinib inevitably
leads to acquired resistance, constraining its therapeutic efficacy in
patients with EGFR‐mutated NSCLC.^[ [50]^2 , [51]^7 , [52]^8 , [53]^9
^] The mechanisms underlying Osimertinib resistance are complex and
varied; thus, precision medicine guidelines necessitate tailoring
treatment regimens based on dynamic changes in resistance patterns,
resulting in increased complexity and limitations within precision
medicine strategies that ultimately contribute to further disease
progression. Therefore, there is an urgent need for broad‐spectrum
therapeutic alternatives to address acquired resistance in malignant
NSCLC.
In our investigation, we demonstrated that Osimertinib‐resistant NSCLC
cells exhibit significantly elevated energy and anabolic demands, along
with exceptionally high levels of oxidation and reduction. Further
analyses revealed a profound dependence on glutamine metabolism.
Glutamine serves multiple roles in cellular metabolism, and glutaminase
(GLS1)‐mediated hydrolysis not only provides energy and reducing
equivalents but also supplies metabolic intermediates as biosynthetic
precursors for the synthesis of non‐essential amino acids, nucleotides,
and fatty acids.^[ [54]^10 ^] Consequently, we contemplated developing
therapeutics targeting glutamine metabolism to refine the treatment
strategy for Osimertinib‐resistant NSCLC.
Recent research has focused on allosteric inhibitors of GLS1 in tumor
glutamine metabolism; however, progress has been limited and clinical
trials have been halted.^[ [55]^11 , [56]^12 ^] Our studies
demonstrated that both the novel macrocyclic GLS1 inhibitor LL202^[
[57]^13 ^] developed by our group and the clinical drug CB839^[ [58]^14
^] did not elicit satisfactory anti‐tumor activity in vivo as
monotherapies. Nevertheless, the present study exposed that concurrent
blockade of the glutamine transporter alanine‐serine‐cysteine type 2
(ASCT2, encoded by SLC1A5) and GLS1 could generate significant
anti‐tumor activity with a manageable safety profile. This dual‐target
approach may elucidate the limited success of selective GLS1 inhibitors
and suggests that a broad‐spectrum strategy could be more effective.
6‐diazo‐5‐oxo‐l‐norleucine (DON) is the most recognized broad‐spectrum
glutamine antagonist; however, it is associated with severe
toxicities.^[ [59]^15 , [60]^16 ^] To address this issue, we leveraged
the high redox level of Osimertinib‐resistant cells to design
antioxidant factor NAD(P)H: quinone oxidoreductase 1 (NQO1)‐responsive
DON prodrug molecules, aiming to reduce toxic side effects while
enhancing therapeutic potency. Our phenotype‐based prodrug molecule
proved more effective than esterase‐ and aminopeptidase‐based prodrug
molecule DRP104 in drug‐resistant cell models.^[ [61]^17 , [62]^18 ,
[63]^19 , [64]^20 ^] In addition to modulating tumor energy metabolism,
NQO1‐responsive prodrug could further augment the anti‐tumor effect by
enhancing the immune microenvironment and activating protective immune
response.
This study offers a novel therapeutic combination for enhancing the in
vivo efficacy of GLS1 allosteric inhibitors targeting glutamine
metabolism. Specifically, the combination of the novel macrocyclic GLS1
inhibitor LL202 and ASCT2 inhibitor V9302 is employed to disrupt both
glutamine uptake and metabolism. The identified NQO1‐responsive DON
prodrugs demonstrate highly promising anti‐tumor activities both in in
vitro and in vivo models of resistance. These findings present a
potential therapeutic alternative for addressing the current challenge
posed by Osimertinib‐resistant NSCLC in clinical settings.
2. Results
2.1. Osimertinib‐Resistant Lung Cancer Cells are Addicted to Glutamine
Metabolism
Given the persistent challenge of secondary resistance to EGFR‐TKI in
treating EGFR‐mutated NSCLC, we established drug‐resistant cell models
(HCC827OR and H1975OR) for Osimertinib, as detailed in the “
Experimental Section”, to investigate targeted treatment strategies.
Enrichment analysis of differentially expressed genes revealed
heightened demands for synthesis and energy metabolism in resistant
cells compared to parental counterparts (Figure [65]1A,B; Figure
[66]S1A,B, Supporting Information). Subsequent deprivation experiments
demonstrated that resistant cells exhibited a greater reliance on
external glutamine rather than glucose for proliferation
(Figure [67]1C; Figure [68]S1C,D, Supporting Information). Further,
glutamine uptake assays confirmed an increased capacity for
^13C‐glutamine uptake in resistant cells (Figure [69]1D). Elevated
levels of GLS1 and ASCT2 (SLC1A5) were also observed in resistant cells
(Figure [70]S1E,F, Supporting Information), with overexpression of
these proteins linked to a higher mortality risk among lung cancer
patients (Figure [71]S1G, Supporting Information). In addition, GLS1
expression was positively correlated with EGFR expression (Figure
[72]S1H, Supporting Information), particularly within the cohort of
EGFR‐mutated lung cancer patients (Figure [73]S1I, Supporting
Information). Collectively, these findings suggest a close association
between reprogrammed glutamine metabolism and upregulation of key
proteins in Osimertinib‐resistant cells.
Figure 1.
Figure 1
[74]Open in a new tab
Reprogramming of glutamine metabolism in cells resistant to
Osimertinib. A,B) KEGG pathway enrichment analysis (A) and GO
functional enrichment analysis (B) of upregulated gene modules in
HCC827OR cells (n = 4 per group). C,E) CCK8 assay showing cell activity
in normal, glucose‐deprived, and glutamine‐deprived cultures (C), and
cell activity following the knockdown of NRF2 (E) (n = per group). D,F)
Quantification of glutamine uptake flux (D) and glutamine and glutamate
levels post NRF2 knockdown (F) using LC‐MS/MS analysis (n = 5 per
group). G,H) Western Blot analysis for proteins in cells treated with
NRF2‐konckdown, with or without the action of MG132 and CQ (G), and
proteins in cells treated with NRF2 inhibitor Brusatol (H). Data are
presented as mean ± SD, calculated using two‐sided unpaired Student's
t‐test, * P < 0.05 and ** P < 0.01.
Moreover, we observed a significant upregulation of NRF2 and its
downstream pathways in resistant cells, which are closely associated
with glutamine metabolism and acquired resistance (Figure [75]S2A,
Supporting Information),^[ [76]^21 ^] as well as poor prognosis in lung
cancer (Figure [77]S2B, Supporting Information). Importantly, our study
revealed that NRF2 activity is not only linked to the proliferative
activity of resistant cells but also to the normal expression of GLS1
protein. Genetic inhibition of NRF2 activity resulted in suppressed
proliferation of resistant cell in both the glutamine‐supplemented and
glutamine‐deprived conditions (Figure [78]1E), along with a decrease in
the conversion of glutamine to glutamate (Figure [79]1F), suggesting
that NRF2 may play a role in GLS1‐mediated glutamine hydrolysis.
Further, inhibition of NRF2 activity led to the activation of cellular
autophagy and significant reduction in GLS1 protein levels without
notable changes at the gene level (Figure [80]1G,H; Figure [81]S2C,D,
Supporting Information). Further assessing autophagy flux using
autophagy inhibitors and autophagy agonists, we demonstrated for the
first time that activated NRF2 maintains intracellular GLS1 protein
levels by inhibiting autophagy‐lysosome function (Figure [82]1G), and
that stable GLS1 expression facilitates orderly hydrolysis of
intracellular glutamine. The key regulator of glutamine metabolism
c‐Myc also influences both glutamine uptake and hydrolysis by
regulating ASCT2 and GLS1 expression.^[ [83]^22 , [84]^23 ^]
Consequently, regulation of glutamine metabolism reprogramming by
glutamine metabolism regulatory genes mainly focuses on the regulation
of key proteins of the glutamine metabolic pathway. In conclusion, we
propose that the dependence on glutamine exhibited by
Osimertinib‐resistant cells is closely related to the functional status
of their metabolic key protein.
2.2. Glutamine Limitation: GLS1 and ASCT2 Function Blocking Contributes to
Osimertinib‐Resistant Lung Cancer Inhibition, Disrupting Energy and Redox
Balance
Subsequently, we explored the potential of targeting glutamine
metabolism by modulating the function of key targets involved in this
pathway to achieve a feasible anti‐tumor effect in the
Osimertinib‐resistant model. First, given the limited in vivo efficacy
of GLS1 inhibitors, we conducted a preliminary exploration and
discovered that the exogenous glutamine uptake significantly increased
following treatment with GLS1 inhibitors in resistant cells (Figure
[85]2A), leading to a certain extent of compensatory glutamine
metabolism. Indeed, glutamine metabolism encompasses not only its
hydrolysis via mitochondrial pathways but also the cellular uptake from
the external environment. Therefore, we contemplated the potential of
targeting both glutamine transport and hydrolysis. Our initial
assessment of inhibitory activity confirmed that co‐targeting of GLS1
inhibitor and V9302 at a 1:1 ratio exhibited significant synergistic
effects in glutamine‐dependent cells (Combination Index [CI] <1) (Table
[86]S1,2, Supporting Information). Further, the synergistic effect of
LL202 (Figure [87]S3A, Supporting Information), a novel macrolide GLS1
inhibitor developed by our group,^[ [88]^13 ^] combined with V9302, was
significantly superior to that observed with CB839 (Figure [89]2B). The
CI value for the LL202 group in resistant cells was below 0.6 (when Fa
> 0.5) (Figure [90]2A; Table [91]S3, Supporting Information),
indicating strong synergy of LL202 and V9302. Similar synergistic
effects were also observed in other glutamine‐dependent lung cancer
cells, with the LL202 combination outperforming CB839 (Figure
[92]S3A–D, Supporting Information). Next, we assessed alterations in
function and phenotype of resistant cells following treatment with
LL202 and V9302 either individually or together. Compared to the
monotherapy groups, combined treatment significantly suppressed cell
proliferation (Figure [93]2C; Figure [94]S4A–D, Supporting Information)
and augmented their capacity to induce apoptosis (Figure [95]2D; Figure
[96]S5A,B, Supporting Information). Moreover, dual‐target inhibition
markedly attenuated invasive potential of resistant cells; while,
reversing epithelial‐mesenchymal transition (EMT) changes
(Figure [97]2E,F; Figure [98]S6A–E, Supporting Information).
Figure 2.
Figure 2
[99]Open in a new tab
GLS1 and ASCT2 function blocking contributes Osimertinib‐resistant
cells inhibition. A) Quantification of glutamine uptake subsequent to
the administration of GLS1 inhibitors to drug‐resistant cells (n = 6
per group). B) Synergistic effect of GLS1 inhibitor with ASCT2
inhibitor on H1975, H1975OR, HCC827, and HCC827OR cells. CI
(combination index) value was calculated as described in the
Experimental Section (n = 10 per group). C–E) CCK8 assay (C) assessing
the cellular activity, flow cytometry analysis (D) for cell apoptosis
and transwell assay (scale bars, 100µm) (E) for cell migration and
invasion, following V9302, LL202, and combined treatment of
Osimertinib‐resistant cells (n = 3 per group). F) The mechanism model
of GLS1 and ASCT2 dual inhibition reversing EMT in
Osimertinib‐resistant cells. Data are presented as mean ± SD,
calculated using two‐sided unpaired Student's t‐test, ** P < 0.01.
The identification of associated metabolic changes revealed a
significant inhibition of the glutamine metabolic pathway resulting
from dual‐target administration (Figure [100]3A; Figure [101]S7D,
Supporting Information). In contrast to single‐target inhibition, the
glycolysis/TCA/oxidative phosphorylation pathways in the combined
treatment group were not activated by stress, indicating that
dual‐target inhibition concurrently disrupts energy metabolism‐related
pathways (Figure [102]3B,C; Figure [103]S7A,B, Supporting Information).
This treatment activated the energy‐regulating switch AMPK, with an
increase in the AMP/ATP ratio observed in the combined group and a
decrease noted in the single‐drug group, suggesting severe energy
stress induced by dual‐target inhibition in tumor cells
(Figure [104]3C; Figure [105]S7C, Supporting Information). Further
analysis of free amino acid levels revealed a significant upregulation
in the combination group (Figure [106]3D), implying enhanced protein
degradation. These findings suggest that dual targeting induces an
energy deficiency within cells, prompting activation of alternative
catabolic pathways to support survival and proliferation.
Figure 3.
Figure 3
[107]Open in a new tab
Combination therapy disrupting energy and redox balance. A,B,D) The
LC/MS‐MS technique was employed to monitor alterations in metabolite
levels within Osimertinib‐resistant cells following administration of
combination therapy. Quantification of Glu/Gln (glutamate/glutamine)
ratio, GSH/GSSG ratio, AMP/ATP ratio (A), heatmap depicting energy
metabolite differences (B), and changes in free amino acids (D) (n = 4
per group). C) Glutamine metabolism inhibition leading to the
activation of distinct metabolic pathways in both monotherapy and
combination therapy. Monotherapy primarily activates energy metabolism,
while combination therapy predominantly induces catabolism and
suppresses anabolism. E) Fluorescence microscopic showing ROS levels of
Osimertinib‐resistant cells after treating with V9302, LL202, and NAC
independently or in combination was determined by DCFH‐DA staining
(scale bars, 100 µm). F) Western Blot analysis for NRF2 and HO‐1
protein following time‐dependent treatment with the combination group.
Data are presented as mean ± SD, calculated using two‐sided unpaired
Student's t‐test. * P < 0.05 and ** P < 0.01.
Given the crucial role of glutathione (GSH), synthesized via the
glutamine metabolic pathway, as a key antioxidant in maintaining redox
homeostasis, we further assessed the cellular redox state. Results from
fluorescence microscopy and flow cytometry demonstrated a significant
increase in intracellular reactive oxygen species (ROS) following the
combined treatment (Figure [108]3E; Figure [109]S7E, Supporting
Information). In addition, administration of the antioxidant
N‐acetylcysteine (NAC) markedly reduced ROS levels in the combination
group and mitigated therapy‐induced cytotoxicity (Figure [110]S7F,
Supporting Information), indicating that ROS generation contributes to
the cytotoxic effects mediated by LL202 and V9302. Further, there was a
significant downregulation of NRF2 and its downstream target HO‐1
following combined treatment (Figure [111]3F), indicating a lack of
reactive activation of the antioxidant pathway. These findings indicate
that dual‐target inhibition enhances protein catabolism for energy
provision while decreasing NRF2 activity within cells (Figure [112]S7G,
Supporting Information). Assessments of autophagic flux using
chloroquine (CQ) as an autophagy inhibitor and MG132 as an autophagy
activator confirmed that dual‐target inhibition suppresses
NRF2/ARE‐mediated antioxidant pathway activation through autophagy
(Figure [113]S7H, Supporting Information). In conclusion, the
dual‐target inhibition of glutamine disrupts energy homeostasis,
concomitantly inducing ROS production and suppressing activation of the
NRF2/ARE antioxidant pathway, thereby further perturbing cellular redox
balance and enhancing anti‐tumor efficacy.
2.3. Combination Therapy Demonstrates Remarkable Efficacy in Preclinical
Models
We evaluated the anti‐tumor efficacy of the combination treatment
across various preclinical models, including 3D tumor cell models,
patient‐derived organoid (PDO) models, and xenograft tumor models. In
the resistant 3D cell model, we observed a dose‐dependent decrease in
cellular activity with increasing concentrations of combination therapy
(Figure [114]S8A–C, Supporting Information). Conversely, activity
decreased in 3D‐HCC827OR compared to 2D cells. To assess clinical
relevance, we established nine lung cancer PDO models using tumor
tissue samples from patients with sequencing information and medication
history (Figure [115]4A,B). In PDOs with EGFR‐activating mutations
resistant to Osimertinib, combination therapy demonstrated greater
inhibition of proliferation than Osimertinib‐resistant 3D cell model
(Figure [116]S8D,E and Table [117]S4, Supporting Information).
Similarly, it inhibited cell proliferation in an ALK‐mutant PDO model
without prior drug exposure. Notably, our LL202 plus V9302 combination
showed equal efficacy in a multi‐drug resistance model that included
chemotherapy, targeted therapy, and immunotherapy. The H1975OR
xenograft model exhibited a significant reduction in tumor, achieving
an impressive 85.18% inhibition rate compared to monotherapy
(Figure [118]4; Figure [119]S9A,B, Supporting Information). Further,
there was a substantial increase in γH2AX expression level and
TUNEL‐positive cells in the combination group (Figure [120]4F,G).
Importantly, no evidence of in vivo toxicity was observed through
monitoring serum ALT/AST and BUN/Cr levels (Figure [121]S9C, Supporting
Information), as well as HE staining of vital organs (Figure [122]S9D,
Supporting Information). In summary, the combined targeted therapy
demonstrated both efficacy and notable safety.
Figure 4.
Figure 4
[123]Open in a new tab
Combination therapy demonstrates superior efficacy in preclinical
models. A) Microscopic showing the morphology of PDOs (scale bars, 500
µm). B) Microscopic showing the morphology of PDOs (scale bars, 100
µm). C–G) H1975OR‐tumor bearing mice were treated with normal saline
(Vehicle), LL202 (30 mg kg^−1), V9302 (30 mg kg^−1), and LL202+V9302
(30 mg kg^−1). Tumors collected at the conclusion of the specified
treatments, representative image depicting the tumors at the end of the
treatment period is presented (C); quantification of tumor weight (D)
and tumor volume (E); representative images of γH2AX positive cells and
TUNEL positive cells in tumors (scale bars, 50 µm) (F), and
quantification of γH2AX positive cells and TUNEL positive cells (G)
(n = 10 per group). Data are presented as mean ± SD, calculated using
two‐sided unpaired Student's t‐test. * P < 0.05 and ** P < 0.01.
2.4. Glutamine Substitution: Design of NQO1‐Activatable DON Prodrugs, and the
Discovery of 10e
The anti‐tumor effect of the LL202 plus V9302 combination treatment was
unsatisfactory in the HCC827OR xenograft model (results not shown),
which is characterized by high energy metabolism. Consequently, we
speculate that dual‐target blockade is insufficient to inhibit in vivo
metabolism in this hypermetabolic tumor model. Does the lack of
efficacy indicate uncertainty regarding glutamine‐targeting strategies?
Notably, the pathways and processes associated with the glutamine
metabolic network are significantly enriched in this model (Figure
[124]5A), suggesting that glutamine metabolism is essential for tumor
progression. Therefore, we propose shifting our antagonistic strategy
toward a broad‐spectrum antagonist capable of comprehensively
inhibiting glutamine metabolism. The most potent broad‐spectrum
antagonist identified in current research is the natural product DON,
which exhibits structural similarities to glutamine, selectively and
irreversibly covalently binding to glutamine‐utilizing enzymes (GUZs).
This binding release the diazo group and forms the enzyme‐DON complex,
effectively inhibiting GUZs activity and disrupting the entire tumor
metabolic network. Although it progressed to phase II clinical trials,
its development has been suspended due to suboptimal tumor targeting
and safety issues leading to gastrointestinal toxicity. As a result,
recent investigations have concentrated on modifying this compound to
enhance its efficacy.^[ [125]^17 , [126]^24 ^]
Figure 5.
Figure 5
[127]Open in a new tab
Design of NQO1‐activatable DON prodrugs. A) Enrichment analysis of
pathway and process closely associated with glutamine in HCC827OR
cells. B) Enrichment analysis of pathways and processes associated with
redox‐reduction. C) GSEA enrichment analysis for GO:0016655
(oxidoreductase activity, acting on NAD(P)H, quinone, or similar
compound as acceptor). D) Differential genes in GO:0016655 (n = 4 per
group). E) Schematic illustration of the design strategy and the
mechanism of action of prodrugs. F) Molecular docking model of NQO1
(PDB ID: 3JSX) with 10e. G) Theoretical mechanism for the liberation of
the bioactive molecule DON from 10e through co‐activation of NQO1 and
esterase.
Further enrichment analysis of the resistance model revealed a
significant enhancement of its antioxidant defense system
(Figure [128]5B). Additionally, key oxidoreductase activities were
markedly elevated, including those associated with NAD(P)H and quinone
receptors (Figure [129]5C). NQO1 is a flavoenzyme that catalyzes
quinone reduction through a two‐electron mechanism in the presence of
NAD(P)H, exerting detoxifying or bioactivating effects. While NQO1 is
constitutively expressed at low levels in normal tissues, it is
overexpressed in tumor tissues and particularly activated in
Osimertinib‐resistant cells (Figure [130]5D; Figure [131]S2A,
Supporting Information). To enhance the safety and efficacy of DON, we
synthesized a novel series of NQO1‐activatable prodrugs for targeting
by leveraging the differential expression between tumor and normal
tissues (Figure [132]5E). Trimethyl lock, a common NQO1 substrate,^[
[133]^25 ^] was used as a protecting group conjugated with the amino
moiety of DON. The carboxyl moiety of DON, meanwhile, was shielded by
esterification, which is similar to DRP104. By modifying substituents
on benzoquinone and introducing diverse ester linkages, we obtained
eight NQO1‐responsive DON prodrugs.
The synthesis and characterization of the designed NQO1‐activatable DON
prodrugs (10a–10h) are illustrated in Scheme [134]S1, Supporting
Information. The catalytic efficacy of prodrugs is detailed in Table
[135]S5, Supporting Information. Prodrug 10e exhibited favorable
metabolic stability and accelerated NQO1‐mediated reduction, leading to
its selection for further assessment. Docking results indicated that
10e specifically bound to the shallow catalytic pocket, which consisted
of Met154 and His161 from one monomer and Tyr126, Tyr128, Met131,
Phe232, and Phe236 from the opposing NQO1, with the benzoquinone
aligned parallel to the cofactor FAD, forming π–π stacking interactions
between both benzene and pyrimidine diketones within the ternary
condensed ring of FAD (Figure [136]5F). Notably, the carbonyl group of
benzoquinone served as a H‐bond acceptor to contact with Tyr126; while,
the ketone and terminal diazo group interacted with FAD through
H‐bonds, providing further insights into the reactivity of 10e toward
NQO1. Subsequently, HPLC analysis was employed to assess the release of
prodrug 10e in the presence of NQO1 and cofactor NADPH; it was observed
that nearly 80% of 10e was depleted within just 10 min when exposed to
NQO1; while, remaining stable in buffer at pH 7.4 in its absence
(Figure [137]S10A, Supporting Information).
Table [138]S5, Supporting Information presents the stability of the
highly metabolized site (plasma) and the toxicity site (intestine).
Under experimental conditions, compound 10e demonstrates considerable
stability in human plasma and intestine, with over 90% of the tested
molecules remaining after 1 h. In rodent plasma, rich in esterase,^[
[139]^17 ^] the ester moiety of the prodrug undergoes hydrolysis into
carboxylic acids. However, the resulting metabolic intermediate,
10e‐int, remains stable in plasma for over an hour (Figure [140]S10B,
Supporting Information). DON is detected only after further
co‐incubation with NQO1 and NADPH, showing a time‐dependent increase
and reaching its maximum release after 2 h (Figure [141]S10C,
Supporting Information). These results indicate that NQO1‐activatable
DON prodrugs exhibit minimal metabolism in both plasma and the
gastrointestinal tract, suggesting a favorable safety profile due to
reduced conversion to DON. Further, in vivo acute toxicity experiments
reveal no significant changes in liver transaminase levels and jejunal
histology in the 10e group compared to DON treatment (Figure
[142]S11A–C, Supporting Information), indicating superior in vivo
safety of 10e relative to DON.
2.5. Tumor Targeting and Anti‐Tumor Activity of 10e
To evaluate the tumor‐targeting properties of 10e, we assessed its
ability to permeate and cleave to DON in tumor cells using tumor‐plasma
model system. 10e exhibited similar DON partitioning and
biotransformation profiles as parental DON drug in tumor cells; while
causing minimal release of DON in human plasma (Figure [143]S12A,
Supporting Information). Further, the distribution ratio of 10e between
tumor cells and plasma was significantly higher than that of DON
itself. Importantly, in the in vivo tumor model, 10e released a greater
amount of DON in tumor tissue than plasma (Figure [144]S12B, Supporting
Information). Metabolite identification revealed that 10e was directly
reduced in tumor cells to form lactone and DON isopropyl ester
(10e‐int2), which was subsequently hydrolyzed to form DON
(Figure [145]5C; Figure [146]S12C, Supporting Information).
Next, we conducted a comparative analysis of the efficacy and safety of
DON, DRP104, and 10e in C57BL/6J mice bearing LLC tumors. The mice
received daily injections of vehicle, DON (0.5 mg kg^−1), DRP104 (0.5
mg kg^−1 DON equiv.), and 10e (0.5 mg kg^−1 DON equiv.), respectively.
Following 6 days of continuous subcutaneous administration (Figure
[147]6A), 10e exhibited comparable antitumor efficacy to both DON and
DRP104. Further, this efficacy was sustained even after discontinuation
(Figure [148]6B,C; Figure [149]S13A,C, Supporting Information).
Although all groups experienced weight loss during treatment, the mice
in the DRP104 and 10e groups showed significant recovery to normal
levels during the discontinuation phase (Figure [150]6D; Figure
[151]S13B, Supporting Information). In addition, gastrointestinal
histological results indicated improvements in both the 10e group and
the DRP104 group compared to the DON group (Figure [152]6F). Notably,
three mice in the DON group died on day 12 of treatment, and one mouse
in the DRP104 group died on day 8 of treatment; conversely, all mice in
the 10e group survived until the experiment's endpoint
(Figure [153]6E). Further, in a murine model of MC38 tumor, 10e (0.5 mg
kg^−1 DON equiv.) demonstrated superior antitumor efficacy
(Figure [154]6G–I; Figure [155]S13D–H, Supporting Information). After 6
days of continuous administration, tumor inhibition rate reached up to
90% with persistent anti‐tumor effects observed until experimental
endpoints following cessation of administration. Despite initial weight
loss during treatment, it returned to baseline levels at later stages
(Figure [156]S13G,H, Supporting Information); while, gastroenteromics
analysis indicated negligible systemic toxicity associated with 10e
throughout treatment (Figure [157]S13F, Supporting Information). These
findings suggest that the in vivo antitumor efficacy of 10e is
comparable to DRP104 and DON but exhibits a significantly improved
safety profile relative to DON.
Figure 6.
Figure 6
[158]Open in a new tab
In vivo evaluation of the antitumor efficacy and safety profile of
optimized prodrug 10e. A–F) Comparison of the efficacy and toxicity of
equimolar DON, DRP104, and 10e (0.5 mg kg^−1 DON equiv.) in LLC
tumor‐bearing mice. C57BL/6J mice were inoculated with 5 × 10^5 cells,
group administration according to the time points shown in the
schematic diagram (A), quantification of tumor volume (B), TGI (C),
body weight (D), and percent survival (E) during the specified
treatments period, and HE staining of the intestines (original
magnification, 20×) (F) at the conclusion of the specified treatments
(n = 6 per group). G–I) The efficacy and toxicity of 10e (0.5 mg kg^−1
DON equiv.) in MC38 tumor‐bearing mice. Quantification of tumor volume
(G) and TGI (H) during the specified treatments period and tumor weight
(I) at the conclusion of the specified treatments (n = 5 per group).
Data are presented as mean ± SD, calculated using two‐sided unpaired
Student's t‐test. * P < 0.05 and ** P < 0.01. Data in (E) were
calculated using log‐rank test.
2.6. Prominent Anti‐Tumor Efficacy of 10e in Immunodeficiency Models with
NQO1 Hyperactivation
We subsequently assessed the anti‐tumor efficacy of 10e in the
NQO1‐hyperactivated HCC827OR transplantation tumor model. In the
subcutaneous transplantation tumor model, 10e (0.4 mg kg^−1 DON equiv.)
and DRP104 (0.4 mg kg^−1 DON equiv.) were administered via subcutaneous
injection for 5 days per cycle over 4 cycles (Figure [159]7A). During
the initial treatment cycle, all mice experienced ≈15% weight loss. In
subsequent cycles, weight loss did not exceed 10%, and the mice
recovered to above normal levels during withdrawal periods
(Figure [160]7F; Figure [161]S14A, Supporting Information).
Gastroenteromics also showed normal morphology (Figure [162]7G). These
findings indicate minimal systemic toxicity in immunodeficient mice
across both treatment cohorts. At the experimental endpoint, the tumor
suppression rate in the 10e group reached 94.25%, significantly higher
than the 66.48% observed in DRP014 group (Figure [163]7E). In addition,
the relative tumor volume (RTV) was 0.805 in the 10e group compared to
4.421 in the DRP104 group, demonstrating markedly superior inhibition
of tumor growth by 10e (Figure [164]7B–E; Figure [165]S14B, Supporting
Information). What's more, significant tumor growth was detected during
the fourth treatment cycle in the DRP104 group; while, no substantial
change was observed in the 10e group, suggesting that DRP104 may lead
to earlier development of resistance. These results confirm the
prominent anti‐tumor efficacy of 10e.
Figure 7.
Figure 7
[166]Open in a new tab
Comparison of the efficacy, toxicity, and metabolism/flow of equimolar
DRP104 and 10e in the NQO1‐activated hypermetabolic tumor model. A–G)
Comparison of the efficacy and toxicity of equimolar DRP104 and 10e
(0.4 mg kg^−1 DON equiv.) in HCC827OR tumor‐bearing mice. BALB/c‐nude
mice were inoculated with 5 × 10^6 cells, group administration was
carried out in accordance with the time points shown in the schematic
diagram (A); tumors collected at the conclusion of the specified
treatments, representative image depicting the tumors at the end of the
treatment period is presented (C); quantification of tumor volume (B);
TGI (E); and body weight (F) throughout the specified treatments
period, as well as measurement of final tumor weight (D) and HE
staining of the intestines (original magnification, 20×) (G) conducted
at the conclusion of the specified treatments (n = 8 per group). H–J)
The LC‐MS/MS technique was employed to monitor alterations in
metabolite levels within HCC827OR tumors at the conclusion of DRP104
and 10e treatment. Comparison of glutamine metabolism and TCA cycle
components (H), KEGG enrichment pathway map (10e vs. DRP104) (I),
differential abundance score (DAS) plots of all enriched metabolic
pathways (The DAS represents the comprehensive change in all
metabolites within the metabolic pathway. A score of 1 indicates
up‐regulation of all identified metabolites in the pathway, while a
score of −1 indicates down‐regulation) (10e vs. DRP104) (J) (n = 4 per
group). Data are presented as mean ± SD, calculated using two‐sided
unpaired Student's t‐test. * P < 0.05, ** P < 0.01, ^# P < 0.05, ^## P
< 0.01, ns not significant, and P > 0.05.
Subsequently, we examined the effects of DRP104 and 10e treatments on
tumor metabolism. Both treatments significantly suppressed tumor
glutamine metabolism, as evidenced by increased glutamine accumulation
and decreased levels of metabolites such as glutamate and GSH, along
with reduced levels of TCA cycle intermediates derived from glutamine
including fumarate, citrate, aspartate, and malate (Figure [167]7H).
Notably, 10e induced more pronounced alterations in these metabolite
levels compared to DRP104. A comprehensive analysis of other key energy
metabolites and a differential analysis of the detected metabolites
were performed (Figure [168]7I,J; Figure [169]S15A–G, Supporting
Information). Treatment with DRP104 resulted in the downregulation of
certain energy metabolite (Figure [170]S15A, Supporting Information);
while, 10e led to downregulation of a greater number of energy
metabolites (Figure [171]S15B, Supporting Information), with
significant differences observed in the differential metabolite
profiles between 10e and DRP104 groups (Figure [172]S15C, Supporting
Information). In addition, KEGG differential enrichment analysis
revealed that compared to DRP104, 10e induced more extensive
downregulation of metabolic pathways, including amino acid metabolism,
nucleotide metabolism, and carbon metabolism, as well as signaling
pathways such as mTOR and PI3K‐AKT (Figure [173]7I,J; Figure
[174]S15D‐G, Supporting Information). These results demonstrate that
10e exerts broad inhibitory effects on tumor metabolism, underscoring
its anti‐tumor advantages, particularly in highly metabolic tumors.
2.7. Compound 10e Promotes Anti‐Tumor Microenvironment in
Immunosuppressive/Immunoinvasive Tumors
Through a comprehensive investigation of the HCC827OR model, we
observed a reduction in the activation and positive regulation of the
immune response (Figure [175]S16A, Supporting Information), as well as
the absence of PD‐L1 immune checkpoint expression (Figure [176]S16C,D,
Supporting Information). Co‐culture experiments with immune cells
revealed that PD‐L1‐deficient HCC827OR cells induced immune cell
infiltration, but without detectable tumor reactivity or significant
tumor cell apoptosis in infiltrated immune cells (Figures [177]S17A–D
and [178]S18A–C, Supporting Information), indicating an immunologically
“cold” microenvironment and challenging prospects for immunotherapy.^[
[179]^26 , [180]^27 ^] Further, there was enhanced regulation of
M2‐type macrophage activation (Figure [181]S16B, Supporting
Information),^[ [182]^28 ^] leading to a significant shift from M0 to
M2 macrophages and promoting a pro‐tumor environment (Figure [183]S19A,
Supporting Information).
Next, we investigated the potential of 10e to modulate the suppressive
immune microenvironment. Following 10e treatment in a co‐culture
system, HCC827OR cell morphology exhibited significant alterations,
including compromised cell membrane integrity and irregular shape
(Figure [184]S18A,B, Supporting Information). Analysis of lysogenic
mediators in the co‐culture environment revealed a marked increase in
levels of anti‐tumor cytokines IL‐2, IL‐12, IFN‐γ, and TNF‐α,
accompanied by significant down‐regulation of pro‐tumor cytokines IL‐4,
IL‐6, and IL‐10 (Figure [185]S18C, Supporting Information). Further,
treatment with 10e induced activation of cytotoxic T lymphocytes such
as CD3^+T, CD4^+T, and CD8^+T cells (Figure [186]8A; Figure [187]S20A,
Supporting Information). Evaluation of tumor‐associated macrophages
based on M1 (CD86^hiiNOS^hi), M2 (CD206^hiARG1^hi) typing, as well as
M1/M2 ratios (Figure [188]8B; Figures [189]S19A and [190]S20B,
Supporting Information) demonstrated that 10e inhibited the
polarization of M2‐type macrophages; while, promoting their
transformation into M1‐type macrophages. Notably, there was a
significant up‐regulation of HLA‐I (HLA‐A, HLA‐B, and HLA‐C) antigens
following exposure to 10e (Figure [191]S19B, Supporting Information),
which is crucial given that loss of HLA class I genotype can render
tumors unresponsive to immunotherapies due to evasion from
immunosurveillance mechanisms.^[ [192]^29 ^]
Figure 8.
Figure 8
[193]Open in a new tab
10e treatment promotes an anti‐tumor microenvironment in
immunosuppressive/ immunoinvasive tumors. A,B) Flow cytometry analysis
was performed to identify immune cell subpopulations following
co‐culture of HCC827OR cells with immune cells, and quantification of
cytotoxic T cells (CD3^+T, CD4^+T, and CD8^+T) (A) and macrophages (M1,
M2, and M1/M2 ratio) (B) (n = 3 per group). C,D) MC38‐bearing mice
treated with 10e (0.5 mg kg^−1 DON equiv., subcutaneously, 6 days),
representative flow cytometry plots (C), and data charts (D) showing
CD3^+T, CD4^+T, CD8^+T, NK, and NKT cells subsets and ratios (n = 3 per
group). E,F) The ELISA assay was employed to monitor alterations in key
cytokines within tumors (E) and spleens (F) of MC38 tumor‐bearing mice
following treatment with 10e (n = 4 per group). Data are presented as
mean ± SD, calculated using two‐sided unpaired Student's t‐test, * P <
0.05, ** P < 0.01, ns not significant, and P > 0.05.
To further elucidate the immunoregulatory mechanism of 10e in
immune‐infiltrating tumors, we conducted immunocyte typing analysis on
MC38 mice treated with 10e (0.5 mg kg^−1 DON equiv.). Consistent with
the findings regarding 10e in immunosuppressive tumors, a significant
disparity in immunocyte typing was observed between treated and
untreated mice. Flow cytometry results revealed that 10e treatment
markedly elevated the proportion of cytotoxic killer immune cells,
including CD3^+T, CD4^+T, CD8^+T, NKT, and NK (Figure [194]8C,D), and
demonstrated heightened proliferative activity of CD4^+T, CD8^+T, and
NKT in the treated group (Figure [195]S21A,B, Supporting Information).
Notably, a substantial upregulation of effector memory T cell subsets
within the population of CD8^+ memory T cells was also observed (Figure
[196]S22A,B, Supporting Information). In contrast to central memory T
cells primarily situated in lymphoid organs, effector memory T cells
are predominantly located in peripheral tissues and can promptly
generate effector cytokines upon antigenic stimulation. ELISA results
demonstrated a significant increase in the levels of anti‐tumor
cytokines IFN‐γ, TNF‐α, and IL‐2 following treatment with 10e
(Figure [197]8E,F). Further, no notable alteration was observed in
immunosuppressive regulatory T (Treg) cell populations (Figure
[198]S22C, Supporting Information); while, there was a marked reduction
in pro‐tumor myeloid‐derived suppressor cell (MDSC) content (Figure
[199]S22D, Supporting Information), suggesting suppression of pro‐tumor
immune response. In summary, the aforementioned results indicate that
10e can substantially enhance the tumor immune microenvironment and
bolster anti‐tumor immune response.
3. Discussion
Our findings imply that targeting glutamine metabolism may provide a
viable therapeutic alternative for Osimertinib‐resistant NSCLC.
Conventionally, strategies to address acquired resistance in NSCLC have
focused on precision medicine, tailoring treatment regimens based on
the evolving nature of resistance. While somewhat effective, this
strategy frequently disregards the metabolic adaptations of cancer
cells and fails to tackle the fundamental cause of the issue of
addressing acquired resistance. Indeed, glutamine assumes a crucial
role in the metabolism of malignant tumors, influencing tumor
progression through interactions with the tumor microenvironment and
contributing to drug resistance and recurrence.^[ [200]^30 , [201]^31 ,
[202]^32 ^]
Our research manifestly demonstrated that the Osimertinib‐resistant
NSCLC model manifests an exceedingly high anabolic activity and energy
demands, presenting a distinct reliance on glutamine metabolism. We
also detected the upregulation of GLS1 and ASCT2, crucial proteins in
glutamine metabolism, within the models. Further, we debuted the
presentation that NRF2,^[ [203]^21 , [204]^32 ^] a regulator intimately
correlated with glutamine metabolism, can modulate glutamine metabolism
by restraining autophagic lysosomal function to preserve GLS1 protein
levels. This finding aligns with the regulatory mechanism of c‐Myc,
which governs glutamine uptake and hydrolysis through modulation of
ASCT2 and GLS1 expression levels.^[ [205]^22 , [206]^23 ^] Our findings
accentuated that the glutamine dependence in Osimertinib‐resistant
cells is intimately affiliated with the functional status of proteins
associated with the glutamine metabolic pathway. Consequently, we
initially contemplated targeting glutamine metabolic pathway‐related
targets.
It is a given that recent studies on tumor glutamine metabolism have
concentrated on the development of GLS1 allosteric inhibitors, to which
our research group has contributed several works.^[ [207]^13 , [208]^33
^] These inhibitors exhibit high selectivity and safety and reveal
promising antitumor activities in vitro; however, achieving limited
progress in vivo and in the clinical setting. Therefore, the current
study investigates the inconsistent activities of GLS1 inhibitors
observed in in vitro and in vivo studies and proposes a novel approach:
simultaneous targeting of tumor glutamine transport (ASCT2) and
glutamine metabolism (GLS1). We hypothesize that this dual‐targeting
strategy might generate a synergistic effect, significantly enhancing
anti‐tumor efficacy (1 + 1 > 2). Our initial investigations confirm the
synergistic benefits of combining LL202 and V9302 over the combination
of CB839 and V9302, as evidenced by cellular activity assays and CI
values. Through comprehensive evaluations both in vitro and preclinical
models, we determine that the LL202 and V9302 combination effectively
inhibits tumor cell proliferation, promotes apoptosis, reduces
invasiveness, and induces energy depletion along with oxidative stress.
This combination therapy also proves effective across various
drug‐resistant organoid models, including those resistant to targeted
therapy, immunotherapy, and chemotherapy. Moreover, in
Osimertinib‐resistant transplanted tumor models, this combination not
only demonstrated significant efficacy but also maintained an excellent
safety profile. Consequently, our study validates and substantiates the
feasibility of this novel dual‐targeting approach to enhance the in
vivo efficacy of GLS1 allosteric inhibitors.
Based on the concept of enhanced efficacy via dual targeting of GLS1
and ASCT2, we further explored if comprehensive inhibition of all GUZs
or broad‐spectrum glutamine metabolism antagonists would surpass the
efficacy of selective GLS1 inhibitors in a clinical context. Distinct
from the DON prodrug design concept of DRP104 relying on Cathepsin L
hydrolysis,^[ [209]^17 , [210]^18 , [211]^19 , [212]^20 ^] this study
further proposed to capitalize on the elevated redox level in the
oxidative reduction within Osimertinib‐resistant cell lines and
designed the antioxidant factor NQO1‐responsive DON prodrug molecules.
This design concept aimed to leverage the unique metabolic and genetic
milieu of cancer cells to enhance therapeutic precision and efficacy.
Through a comprehensive assessment of stability, tumor targeting, and
safety, we identified the optimal NQO1‐responsive DON prodrug, 10e, a
molecule that demonstrated potent antitumor activity in diverse ex vivo
and in vivo scenarios and exceeded the efficacy of the clinical phase
II drug, DRP104, in the Osimertinib‐resistant malignant tumor model.
Given the potent antitumor effect and the vigorous metabolism of
Osimertinib‐resistant cells in tumors, we further evaluated its impact
on tumor metabolism. The results indicated that 10e not only disrupted
the glutamine metabolic network but also partially influenced glucose
and lipid metabolism in tumor tissues, and the effect was more
prominent than that of DRP104. Hence, our results imply that 10e holds
a therapeutic advantage over DRP104 in hypermetabolic tumors.
The interactions between tumor cells and their microenvironment are of
critical significance for the cancer phenotype, disease progression,
and therapeutic response.^[ [213]^34 , [214]^35 ^] DON and its
prodrugs, JHU083 and DRP104, have been demonstrated to be efficient in
improving the tumor microenvironment, thereby generating superior
antitumor activity.^[ [215]^16 , [216]^17 , [217]^24 ^] Our findings
also suggest that 10e activates cytotoxic immune cells and elicits a
protective immune response. Currently, the modulation of the tumor
immune microenvironment by DON prodrugs is mainly centered on
immune‐infiltrating tumor models,^[ [218]^17 , [219]^24 ^] while in
this investigation, we not only explored immune‐infiltrating tumors but
also concentrated on the effect of 10e on the Osimertinib‐resistant
model of an immunologically “cold” microenvironment. Metabolic
reprogramming, coupled with PD‐L1 deletion, elevated levels of IL‐6,
and extensive infiltration of tumor‐associated macrophages results in a
cold immune phenotype in Osimertinib‐resistant models, which
complicates the efficacy of immunotherapy.^[ [220]^26 , [221]^36 ,
[222]^37 , [223]^38 , [224]^39 ^] Interestingly, our results reveal
that 10e not only promotes the activation of cytotoxic killer immune
cells in this tumor microenvironment but also reverses the infiltration
proportion of M2/M1‐type macrophages. Thus, our research herein could
manifest that 10e notably ameliorates the “cold” immune
microenvironment and potentiates the anti‐tumor immune response,
thereby potentially resolving the immunotherapy conundrum.
4. Conclusion
In summary, our findings reinforce the concept of targeting rewired
metabolism in cancer, particularly in addressing Osimertinib‐resistant
NSCLC. The compound 10e demonstrates significant promise in the
context, highlighting the potential to achieve a favorable therapeutic
index. Given the complexity of glutamine metabolism in cancer, a
comprehensive inhibition strategy that targets multiple nodes, such as
ASCT2 and GLS1, or a broad inhibition is essential to prevent rapid
metabolic rewiring. Mechanism‐based prodrug design holds the potential
to expand the therapeutic window and enhance the clinical application
of anti‐tumor drugs targeting tumor glutamine metabolism.
5. Experimental Section
Chemical Synthesis
Details of the synthesis of the compounds and their intermediates are
elaborated in the Supporting Information.
Reagents
LL202, 10e were obtained from our laboratory with a purity > 98%. DON
(HY‐108357), DRP104 (HY‐132832), CB‐839 (HY‐12248), V9302 (HY‐112683),
NAC (HY‐B0215), Brusatol (HY‐19543), Chloroquine (HY‐17589A),
Osimertinib (HY‐15772), MG‐132 (HY‐13259), L‐Glutamine (HY‐N0390),
L‐Glutamic acid (HY‐14608), and GSH (HY‐D0187) were obtained from MCE.
^13C‐L‐Glutamine (IR‐73028) was obtained from Shanghai Zzbio.
L‐Glutamine solution ([225]R27053) was obtained from Shanghai yuanye
Bio‐Technology. NADPH ([226]BD123569) was obtained from Shanghai Bide
Pharmatech.
Cell Culture
16HBE, A549, H596, H1299, PC9, H1975, HCC827, LLC, and MC38 cells were
purchased from the National Collection of Authenticated Cell Cultures.
H1975OR and HCC827OR were self‐induced Osimertinib‐resistant cells in
the authors’ laboratory from H1975 and HCC827 by increasing initial
concentration(IC[50] value) gradually. All cells were cultured under
standard culture conditions (37 °C, 5% CO[2]) in the culture medium
recommended by the ATCC. None of the cells were contaminated by
mycoplasma.
Tumor Cell‐Immune Cell Co‐Culture
With respect to T‐cell activation, Anti‐CD3 antibody was diluted in
sterile PBS (1 µg mL^−1), and the diluted antibody was introduced into
a T25 culture flask. The flasks were incubated at 5% CO[2], 37 °C for 2
h. Then, the anti‐CD3 solution was aspirated, and the Jurkat T cells
were added to the CD3‐activated flasks concurrently with 5 µg mL^−1 of
anti‐CD28; the Jurkat T cells were incubated at 5% CO[2], 37 °C for
48–72 h.
The tumor cells were inoculated in 6‐well plates overnight, and
subsequently, co‐cultured directly with pre‐activated Jurkat T cells at
a 1:1 ratio. After 72 h, the tumor cells were collected and analyzed
for apoptosis using Annexin V‐FITC/PI in a flow cytometer.
Pre‐inoculated tumor cells were directly co‐cultured with pre‐activated
Jurkat T cells at a 1:1 ratio for 72 h (treated or untreated with 10e),
following which, the cell morphology was observed under the microscope
(Nikon ECLIPSE Ts2). Eventually, T cells were collected and qPCR was
carried out to detect the levels of relevant cytokines, respectively.
Fresh mouse spleens were harvested in an aseptic setting, adding 5 mL
of sterile PBS, ground, and filtered. The cells were centrifuged at
1000 rpm for 5 min at 4 °C, the supernatant was disposed of, and the
cells were resuspended in 2–5 mL of pre‐chilled 1× RBC lysate for 5
min. Then, the cells were resuspended in 10 mL PBS, centrifuged at 1000
rpm for 5 min at 4 °C, after which the cells were re‐suspended in
PRMI‐1640 medium. Next, the suspended splenocytes were directly
co‐cultured with pre‐inoculated tumor cells. After 72h of culturing
(treated or untreated with 10e), splenocytes were collected and
assessed by flow cytometry (FACSCelesta) for immune cell typing.
THP‐1 cells were inoculated into 6‐well plates and treated with 100 ng
mL^−1 PMA. The cells were incubated at 5% CO[2] and 37 °C for 24 h.
After incubation, the supernatant was discarded and replaced with
PRMI‐1640 medium. Once the cells reached ≈80% confluence, culture
chambers were placed in the 6‐well plates, and tumor cells were added
to the upper layer. The culture was continued for 48–72 h with or
without 10e treatment. Subsequently, macrophages from the lower layer
were collected. qPCR was employed to detect macrophage markers and
cytokines, and flow cytometry (FACSCelesta) was utilized to analyze
macrophage typing.
Western Blot
Cell samples were lyzed using RIPA buffer supplemented with protein
phosphatase inhibitor and protease inhibitor. The lysates were
separated by SDS‐PAGE and transferred to polyvinylidene fluoride (PVDF)
membrane. The membrane was blocked with 5% nonfat milk, and then,
incubated with diluted primary antibodies at 4 °C for 12 h and washed
with 1× TBST; the membrane was incubated with a secondary antibody for
1 h at room temperature. The immune response bands were visualized
using enhanced chemiluminescent reagents (Tanon). The primary
antibodies used are listed in Table [227]S7, Supporting Information.
Quantitative Real‐Time PCR
Total ribonucleic acid (RNA) was extracted from cells using RNA
isolater Total RNA Extraction Reagent (Vazyme) according to the
manufacturer's instructions. RNA concentration and purity were measured
by One Drop spectrophotometer (Wuyi Technology) to ensure the quality
of RNA. RNA was reverse transcribed into complementary deoxyribonucleic
acid (cDNA) using the Hiscript Q RT SuperMix for qPCR (+gDNA wiper) Kit
(Vazyme). Quantitative analysis was performed using Taq pro Universal
SYBR qPCR Master Mix Kit (Vazyme) and QuantStudio 3 real‐time PCR
instrument (Thermo Fisher Scientific). Relative mRNA levels were
standardized to the β‐actin mRNA levels. The primer sequence
information is listed in Table [228]S8, Supporting Information.
Transcriptomics
Total RNA samples were extracted using an RNA Extraction Kit (Takara)
according to the manufacturer's instructions. Total amounts and
integrity of RNA were assessed using the RNA Nano 6000 Assay Kit of the
Bioanalyzer 2100 system (Agilent Technologies). Libraries were
sequenced on an Illumina HiSeq 6000 platform. After quality control of
raw reads, the clean reads were mapped to the human genome using
default parameters. Differentially expressed gene (DEG)–normalized read
counts (fragments per kilobase of exon per million [FPKM]) were
calculated using feature Count (v1.5.0‐p3). The clusterProfiler R
package (3.8.1) was used to test the statistical enrichment of
differentially expressed genes in the Reactome pathway, Kyoto
Encyclopedia of Genes and Genomes (KEGG) pathways, and Gene Ontology
(GO) pathway. KEGG pathways and GO were annotated using the KEGG
pathway database ([229]http://www.genome.jp/kegg/) and GO Database
([230]http://www.geneontology.org/), respectively.
Liquid Chromatography–Mass Spectrometry/Mass Spectrometry (LC‐MS/MS)
LC‐MS/MS system (SHIMADZU) was used to measure the ^13C‐L‐glutamine
uptake content and metabolite content in cells and tumor tissue. The
aqueous phase was 0.1% formic acid solution, and the organic phase was
acetonitrile. Cells were washed with ice‐cold PBS, repeatedly
freeze‐thawed with pure water, and lyzed by sonication to prepare cell
samples for amino acid quantification. ^13C‐L‐glutamine was diluted
with assay buffer solution (137 mm NaCl, 5.1 mm KCl, 0.77 mm
KH[2]PO[4], 0.71 mm MgSO[4], 1.1 mm CaCl[2], 10 mm D‐glucose, and 10 mm
HEPES, PH = 6.5) to a final concentration of 10 µm. The cells were
washed with buffer solution (hot) for three times, incubated with
^13C‐L‐glutamine at 37 °C for 15 min, then washed with buffer solution
(cold) for three times, repeatedly freeze‐thawed with pure water, and
lyzed by sonication to prepare cell samples for ^13C‐L‐glutamine uptake
assays. The collected tumor cells were weighed to the same mass and
homogenized in the medium of methanol: water (80:20) to extract the
metabolites, and stored at −80 °C for more than 2 h to precipitate the
proteins, followed by centrifugation at 4 °C at 12 000 rpm for 10 min
to isolate the supernatant. After drying, 50% acetonitrile was added to
re‐dissolve for metabolite content assays. The samples were extracted
with acetonitrile, centrifuged, injected with Shim‐pack Scepter
Diol‐HILIC‐120 (2.1 × 100 mm, 1.9µm, SHIMADZU), and monitored in
positive ion mode on a triple quadrupole 8045 mass spectrometer.
Protein concentrations were determined by processing parallel groups to
standardize amino acids content.
Flow Cytometry Analysis (FACS)
For the analysis of apoptosis, cells were seeded in six‐well plates and
incubated for 12–24 h at 5% CO[2], 37 °C. The compounds were prepared
into a master mix of 10^4 µm with DMSO, and then, diluted to an
appropriate concentration with medium containing 10% FBS. This allowed
the drugs to stimulate the cells for 24–48 h. Cells and cell
supernatants were collected and washed by centrifugation at 1800 rpm,
4 °C for 5 min. Then, 100 µL of 1× Binding Buffer was added for a
single‐cell suspension. After that, 5µL of Annexin V‐FITC and 5 µL of
PI Staining Solution were added to the sample tubes, incubated at room
temperature away from light for 10 min, and after incubation, 400 µL 1×
Binding Buffer was added. The stained samples were detected by flow
cytometry (FACSCelesta) within 1 h.
For the analysis of immune cells, MC38‐bearing mice were administered
subcutaneously for 6 days when the tumor size reached ≈400 mm^3. The
tumors were collected within 1 h after the final dose. Isolated tumor
cells and spleen were washed with PBS; 4 mL of digestive solution was
added to the tumor samples and digested at 37 °C, 100 rpm for 25 min.
After digestion, the tumor cells were ground and washed with PBS and
centrifuged at 2000 rpm for 5 min; the spleen samples were ground and
added to erythrocyte lysate, washed with 4 mL PBS, and centrifuged at
2000 rpm for 5 min. After that, the tumor samples were resuspended with
5 mL 40% Percoll and slowly added to 70% Percoll, followed by
centrifugation at 20 °C, 2000 rpm for 20 min to obtain single nucleated
cells. Then, tumor and spleen mixed samples were blocked with Mouse FcR
blocking reagent, and the antibodies were added to stain samples at
4 °C for 30 min in the dark. The samples to be membrane‐broken were
washed with PBS and centrifuged at 2000 rpm for 5 min, and the cells
were resuspended with fix/perm membrane‐broken solution, and the
membrane was broken at 4 °C for 45 min. After membrane‐braking, the
cells were washed and resuspended with perm/wash membrane‐broken
solution, and then, the corresponding antibodies were added and the
cells were incubated at 4 °C for 30 min in the dark. After incubation,
the cells were washed by adding perm/wash, resuspended with PBS. All of
the stained samples were detected by flow cytometry (FACSCelesta). The
antibodies of flow cytometry analysis used are listed in Table [231]S9,
Supporting Information.
Enzyme Linked‐Immunosorbent Assay (ELISA)
The antigenicity of all samples was evaluated using Polyclonal
Antibodies Quantification Kits purchased from DAKEWE Biotech Co., Ltd.,
China. To each well, 100 µL of standard solutions and samples were
added and incubated with 50 µL Biotinylated Antibody at 37 °C for 90
min. The wells were then emptied and washed four times with 300 µL 1×
Washing Buffer, followed by 100 µL of the Streptavidin‐HRP solution,
incubation at 37 °C for 30 min in the dark, and washing with 1× Washing
Buffer four times. Then, 100 µL of TMB was added and incubated at 37 °C
for 20 min in the dark. The reaction was stopped by adding 100 µL of
stop solution into each well, and the absorbance was measured
simultaneously at measurement wavelength 450 nm and reference
wavelength 630 nm by microplate reader (Molecular Devices).
Organoids Establishment and Drug Sensitivity Test
Patient‐derived fresh tumor tissues were further cultured into PDO
models. H1975OR and HCC827OR cells were collected, diluted with
Matrigel (Corning), and seeded with gel drops (10 µL, 3000 cells) for
3D culture. After the cells could proliferate normally and stably
passaged, the 3D model was successfully constructed. The organoids
(PDOs and 3D models) were photographed and recorded. Organoids were
dissociated into homogeneous cell masses, resuspended, mixed with
Matrigel at a ratio of 2:3 and seeded into 384‐well plates (10 µL, 2000
cells per well). After gelation, medium was added to culture cells for
48 h, then treated with drugs for 96 h. Cell viability was measured
using the CellTiter‐Glo Viability Assay Kit (Promega) according to the
manufacturer's instructions. The basic information of organoids is
listed in Table [232]S4, Supporting Information.
In Vivo Experiments
For single‐dose acute toxicity test, the C57BL/6J (6–7 weeks) mice were
divided into four groups containing Vehicle, DON (6 mg kg^−1, s.c.),
10e (6 mg kg^−1 DON equiv., s.c.), and 10e (12 mg kg^−1 DON equiv.,
s.c.). All the agents were administered by subcutaneous injection for
one time, and then the survival situation was observed.
For LLC subcutaneous tumor model, ≈5 × 10^5 LLC cells suspended in PBS
(100 µL) were injected into the right axilla of C57BL/6J mice (6–7
weeks). After the tumors grew to ≈100 mm^3, all the mice were
randomized into four groups (six mice for each group) containing
Vehicle (normal saline, s.c.), DON (0.5 mg kg^−1, s.c.), DRP104 (0.5 mg
kg^−1 DON equiv., s.c.), and 10e (0.5 mg kg^−1 DON equiv., s.c.), and
all the agents were continuously administered for 6 days on 8 days off
by subcutaneous injection.
For MC38 subcutaneous tumor model, ≈3 ×10^5 MC38 cells suspended in PBS
(100 µL) were injected into the right axilla of C57BL/6J mice (6–7
weeks). After the tumors grew to ≈100 mm3, all the mice were randomized
into two groups (five mice for each group) containing Vehicle (normal
saline, s.c.) and 10e (0.5 mg kg^−1 DON equiv., s.c.), and all agents
were continuously administered for 6 days on 8 days off by subcutaneous
injection.
For H1975OR subcutaneous tumor model, ≈1 × 10^7 H1975OR cells suspended
in PBS (100 µL) were injected into the right axilla of BALB/c‐Nude mice
(5–6 weeks). After the tumors grew to ≈100 mm^3, all the mice were
randomized into four groups (ten mice for each group) containing
Vehicle (normal saline, s.c.), LL202 (30 mg kg^−1, s.c.), V9302 (30 mg
kg^−1, s.c.) and LL202+V9302 (30 mg kg^−1, s.c.), and all agents were
continuously administered for 20 days by subcutaneous injection.
For HCC827OR subcutaneous tumor model, ≈5 × 10^6 HCC827OR cells
suspended in PBS (100 µL) was injected into the right axilla of
BALB/c‐Nude mice (5–6 weeks). After the tumors grew to ≈100 mm^3, all
the mice were randomized into three groups (eight mice for each group)
containing Vehicle (normal saline, s.c.), DRP104 (0.4 mg kg^−1 DON
equiv., s.c.) and 10e (0.4 mg kg^−1 DON equiv., s.c.), and all agents
were continuously administered for 5 days on 4 days off for first
cycle, and continuously administered for 5 days on 2 days off for the
next three cycles by subcutaneous injection.
DRP104 and 10e were dissolved in 5% DMSO and 0.9% saline solution. DON
was diluted to the appropriate concentration with 0.9% saline solution.
The tumor growth and body weights were measured and recorded every day.
As the endpoint of treatment, mice were sacrificed for collecting tumor
and gastrointestinal tissue, and washed with PBS. The gastrointestinal
tissue was taken and fixed by adding 4% paraformaldehyde, and tumors
were weighed and photographed after water removal. Then, samples were
rapidly frozen in liquid nitrogen and stored in −80 °C for histology
and metabolomics. The relevant formula was calculated as:
[MATH: TumorvolumeTV,mm3=lengthmm×widthmm22 :MATH]
(1)
[MATH: TumorgrowthinhibitionTGI,%
=1−TVTreatment_DayN−TVTreatment_Day0TVVehicle_DayN−TVVehicle_Day
0×100% :MATH]
(2)
TV[Treatment_DayN]: tumor volume on the day of administration for the
treatment group, TV[Treatment_Day0]: tumor volume at the beginning of
the experiment for treatment group, TV[Vehicle_DayN]: tumor volume on
the day of the experiment for vehicle group, and TV[Vehicle_Day0]:
tumor volume at the beginning of the experiment for vehicle group.
All animal experiments were performed according to protocols approved
by the Animal Ethics Committee of China Pharmaceutical University
(Approval No.2023‐10‐018, 2024‐05‐024, 2024‐07‐016).
Histology
For Immunohistochemistry (IHC), the heart, liver, spleen, lung, kidney,
and tumor of mice were removed quickly and fixed with 4%
paraformaldehyde for 12 h. The tissue sections were stained using
UltraSensitiveTM SP (Mouse/Rabbit) IHC Kit (Maxim) and anti‐γH2AX
antibody (Santa Cruz Biotechnology) according to manufacturer's
instructions. Tissue sections were stained using the TUNEL Apoptosis
Detection Kit (Keygen Biotech) according to the manufacturer's
instructions. Histological sections were scanned using the Olympus
inverted phase contrast microscope (Olympus Corporation) at 400× field
of view.
For hematoxylin‐eosin staining (H&E), animal models were executed as
mentioned above and dissected surgically for the evaluation of possible
pathological changes. The collection of gastrointestinal tissue was
fixed in 10% buffered formalin, dehydrated in ethanol, embedded in
paraffin, and then stained with hematoxylin and eosin. The pathological
changes were captured with a Nikon 80i optical microscope.
Statistical Analysis
The results are presented as the mean ± SD. Differences between groups
were analyzed for statistical significance using t‐test analysis in
GraphPad Prism (Version 8.3.0, GraphPad Software Inc., LLC).
Statistical significance was accepted with P < 0.05, *: P < 0.05, **: P
< 0.01, ^#: P < 0.05, ^##: P < 0.01, and ns: no significant difference.
The data analysis was conducted by using GraphPad Prism and SPSS
(Version 25, IBM SPSS Software Statistics Inc., LLC).
Conflict of Interest
The authors declare no conflict of interest.
Author Contributions
J.H. and J.B. conceived the project. J.H., X.Z., and H.Z. carried out
the biological experiments and interpreted data. J.H. and Y.L.
synthesized, purified, and characterized compounds. H.H., Z.Q., H.W.,
and X.X. were involved in directing the experiment. Z.L. and D.H.
oversaw and provided insights and materials. J.H. and J.B. wrote the
manuscript with input from all authors. Z.L., D.H., and J.B. provided
supervision.
Supporting information
Supporting Information
[233]ADVS-12-2411479-s001.docx^ (15.2MB, docx)
Acknowledgements