Abstract

   Pancreatic ductal adenocarcinoma (PDAC) is a highly malignant tumor
   with a dismal prognosis. Gemcitabine‐based chemotherapy has emerged as
   a first‐line treatment for PDAC. However, the development of
   gemcitabine resistance often results in therapeutic failure. In order
   to uncover the underlying mechanisms of gemcitabine resistance,
   gemcitabine‐resistant PDAC cell lines and patient‐derived xenograft
   (PDX) models are established and subjected to RNA sequencing. It is
   found that CMTM6 is closely related to gemcitabine resistance in PDAC.
   Multi‐omics analysis revealed that EP300‐mediated H3K27ac modification
   is involved in the transcriptional activation of CMTM6, which maintains
   IGF2BP1 expression by preventing its ubiquitination. The m^6A reader
   IGF2BP1 stabilizes the EP300 and MYC mRNAs by recognizing m^6A
   modifications, forming a positive feedback loop that enhances tumor
   stemness and ultimately contributes to PDAC resistance. The combined
   application of the EP300 inhibitor inobrodib and gemcitabine exerts a
   synergistic effect on PDAC. Overall, these findings reveal that the
   EP300–CMTM6–IGF2BP1 positive feedback loop facilitates gemcitabine
   resistance via epigenetic reprogramming and the combined use of
   inobrodib and gemcitabine represents a promising strategy for
   overcoming chemoresistance in PDAC, warranting further investigation in
   clinical trials.

   Keywords: chemoresistance, CMTM6, epigenetic modification,
   N6‐methyladenosine, pancreatic ductal adenocarcinoma
     __________________________________________________________________

   EP300‐mediated H3K27ac modification upregulates CMTM6, driving
   gemcitabine resistance in PDAC through an EP300‐CMTM6‐IGF2BP1 positive
   feedback loop facilitated by epigenetic reprogramming. This loop
   reinforces tumor stemness by stabilizing MYC and EP300 mRNAs via
   IGF2BP1‐mediated m6A modification. Combined treatment with the EP300
   inhibitor Inobrodib and gemcitabine demonstrates a synergistic effect,
   representing a promising strategy to overcome chemoresistance in PDAC.

   graphic file with name ADVS-11-2406714-g006.jpg

1. Introduction

   Pancreatic ductal adenocarcinoma (PDAC) is one of the most lethal
   malignancies, with a 5‐year survival rate of less than 10%.^[ [50]^1 ^]
   To date, surgery is the only potential cure for PDAC; however, up to
   80% of PDAC patients are diagnosed at an advanced stage when they are
   no longer eligible for curative therapies.^[ [51]^2 ^] Chemotherapy is
   a primary treatment for PDAC, and gemcitabine (GEM) is a crucial drug
   in its management.^[ [52]^3 ^] However, the overall response rate of
   PDAC patients to GEM remains less than 20%.^[ [53]^4 ^] This is
   attributed to the development of GEM resistance, leading to treatment
   failure in PDAC. Thus, exploring the mechanism of GEM resistance in
   PDAC is of great value for finding new therapeutic targets and
   improving the prognosis of PDAC patients.

   CMTM6 belongs to the chemokine‐like factor‐like MARVEL transmembrane
   domain‐containing family (CMTM)^[ [54]^5 ^] and is mainly recognized as
   a critical regulator of programmed cell death 1 ligand 1 (PD‐L1).^[
   [55]^6 ^] CMTM6 regulates PD‐L1 by inhibiting its ubiquitination,^[
   [56]^6 ^] thus preventing its lysosome‐mediated degradation in
   recycling endosomes.^[ [57]^7 ^] Previous studies have focused on the
   effect of CMTM6 on antitumor immunity,^[ [58]^8 ^] and its broader role
   in tumor development and drug resistance has been recently revealed.
   CMTM6 is highly expressed in gliomas and indicates poor prognosis.^[
   [59]^9 ^] It also enhances tumor stemness in head and neck squamous
   cell carcinoma.^[ [60]^10 ^] Regarding chemoresistance, CMTM6 promotes
   cisplatin resistance by regulating the Wnt signal pathway through the
   ENO‐1/AKT/GSK3β^[ [61]^11 ^] and AKT/c‐Myc axes.^[ [62]^12 ^] CMTM6
   could repress p21 ubiquitination to regulate chemoresistance in
   hepatocellular carcinoma.^[ [63]^13 ^] Nevertheless, the precise role
   of CMTM6 in the chemotherapy resistance of PDAC remains unclear.

   Epigenetic reprogramming is a distinctive feature of malignant
   tumors.^[ [64]^14 ^] Epigenetics is a field of study that explores
   stable phenotypic changes independent of DNA sequence, including
   histone modifications, m^6A modifications, and other epigenetic
   phenomena. Histone modification is a covalent post‐translational
   modification of histones that modulates DNA‐based processes and
   includes methylation, acetylation, phosphorylation, ubiquitylation,
   etc. A common histone modification, histone H3K27 acetylation
   (H3K27ac), which involves the acetylation of the lysine residue at the
   N‐terminal position 27 of the histone H3 protein, is a marker for gene
   transcriptional activation and a driver of tumor development.^[ [65]^15
   ^] The CREB‐binding protein (CEBBP/CBP) and P300 (EP300/P300) are the
   main histone acetyltransferases overexpressed in multi‐tumor and
   drug‐resistant cells. Thus, targeting EP300 has emerged as a potential
   therapeutic approach for tumors. Two EP300 inhibitors, inobrodib
   (CCS1477) and FT‐7051, are currently under clinical evaluation in
   patients with advanced and drug‐resistant solid tumors or hematological
   malignancies.^[ [66]^16 ^] Preclinical and early‐phase clinical data
   have revealed the potential of inobrodib in treating hematologic
   malignancies.^[ [67]^17 ^] However, there is limited research regarding
   the role of epigenetic reprogramming in GEM‐resistant PDAC.

   N6‐methyladenosine (m^6A) modification is the most abundant mRNA
   modification in eukaryotes and plays a critical role in various
   fundamental biological processes. It is dynamically regulated by
   methylases (writers) and demethylases (erasers). The specific proteins
   that recognize m^6A modifications and determine the fate of bound RNA
   are known as m^6A readers. IGF2BP1 is a newly characterized m^6A reader
   belonging to the insulin‐like growth factor 2 mRNA‐binding protein
   (IGF2BP) family.^[ [68]^18 ^] IGF2BP1 has been recently reported to
   enhance tumor stemness by mediating the stabilization of βTrCP1 and
   c‐MYC mRNAs in response to β‐catenin signaling.^[ [69]^19 ^]

   In this study, we aimed to investigate the molecular mechanism
   underlying CMTM6 upregulation in PDAC and its influence on GEM
   resistance. Our findings present a novel therapeutic target for
   GEM‐resistant PDAC.

2. Results

2.1. CMTM6 Is Upregulated in GEM‐Resistant PDAC

   To identify genes associated with GEM resistance, we established a
   patient‐derived xenograft (PDX) model using PDAC tissues. Following
   three generations of gemcitabine treatment (F3–F5), the tumor growth
   inhibition in the GEM‐resistant PDX models decreased significantly
   compared to the wild‐type (WT) PDX models, confirming the development
   of resistance to GEM (Figure [70]1A; Figure [71]S1A, Supporting
   Information). GEM‐resistant PDAC cell lines were established using
   BxPC‐3 and CFPAC‐1 cells and certified by IC50 value (Figure [72]s1 B,
   Supporting Information). Then, we performed RNA‐sequencing on tissues
   from both WT and GEM‐resistant PDX (Figure [73]1B), as well as on WT
   and GEM‐resistant PDAC cells.

Figure 1.

   Figure 1
   [74]Open in a new tab

   CMTM6 is upregulated in GEM‐resistant PDAC. A) Schematic representation
   of the establishment of wild‐type (WT) and GEM‐resistant (GR) PDX
   models from PDAC; B) Heatmaps displaying the gene expression profiles
   of WT and GR PDXs; C) Venn diagram depicting upregulated genes in GR
   BxPC‐3 and CFPAC‐1 cells (data from GEO database), GR PDXs (RNA‐seq),
   and PDAC tumor tissues (data from TCGA database); D) Representative
   images of IHC staining for CMTM6 in different passages of WT and GR
   PDXs. Scale Bars, 100 µm; E) Upregulated CMTM6 expression in BxPC‐3/GR
   and CFPAC‐1/GR cells ([75]GSE140077 dataset); F) Elevated CMTM6
   expression in PDAC tumor tissues compared to normal pancreatic tissues
   (TCGA and GTEx datasets); G) Kaplan‐Meier survival curves of PDAC
   patients stratified by CMTM6 expression levels (TCGA datasets); H)
   Dose‐dependent increase in CMTM6 expression in BxPC‐3 cells under GEM
   treatment; I) Time‐dependent increase in CMTM6 expression in BxPC‐3
   cells under GEM treatment; J) Representative CT images illustrating
   GEM‐sensitive (progressive disease, PD) and GEM‐resistant (partial
   response, PR) PDAC tumors before and after GEM‐based chemotherapy
   (left). Multi‐immunofluorescence (mIF) staining for CMTM6 in
   corresponding needle biopsy specimens obtained before chemotherapy
   (right). Scale Bars, 30 µm; K) Statistical analysis of CMTM6 expression
   according to mIF results in PDAC tumors of PD and PR patients. Error
   bars represent the mean ± SD (n = 3 in E, H, and I). ^** p < 0.01; ^***
   p < 0.001 according to Student's t‐test.

   Comprehensive analysis, together with the pancreatic adenocarcinoma
   (PAAD) dataset from The Cancer Genome Atlas (TCGA), revealed that the
   ANTXR1, CMTM6, ITGB6, PIGR, and PTGS2 genes were significantly
   upregulated in GEM‐resistant PDAC cell lines (Figure [76]1C). IHC
   staining confirmed that the level of CMTM6 progressively increased
   across the generation of GEM‐resistant PDX tumors (Figure [77]1D;
   Figure [78]s1 C, Supporting Information). Consistent with our results,
   datasets from the Gene Expression Omnibus (GEO) database showed
   elevated levels of CMTM6 in GEM‐resistant cells (Figure [79]1E).
   Moreover, TCGA datasets (excluding nontumor samples and non‐PDAC
   samples) indicated that high CMTM6 expression in PDAC tumors correlated
   with an unfavorable prognosis (Figure [80]1F,G).

   We treated BxPC‐3 cells with GEM and found that CMTM6 expression
   increased in a time‐ and dose‐dependent manner (Figure [81]1H,I; Figure
   [82]S1D,E, Supporting Information). Needle biopsy specimens from PDAC
   patients were collected prior to GEM‐based chemotherapy, and their
   response to chemotherapy was followed up. Using multiplex
   immunofluorescence (mIF) staining, we observed that CMTM6 was
   significantly elevated in chemotherapy‐resistant tissue specimens
   compared to chemotherapy‐sensitive specimens (Figure [83]1J,K).

2.2. EP300‐Mediated Epigenetic Modification Activates CMTM6 Transcription in
PDAC

   Next, we attempted to uncover the mechanism behind the upregulation of
   CMTM6 in GEM‐resistant PDAC. According to the cBioPortal data, CMTM6
   mutations are very rare in PDAC, which could not fully explain its
   significant overexpression (Figure [84] 2A). Therefore, we speculated
   that increased CMTM6 expression might be related to nonmutational
   epigenetic modification.

Figure 2.

   Figure 2
   [85]Open in a new tab

   EP300‐mediated epigenetic modification activates CMTM6 transcription in
   PDAC. A) Mutation frequency of CMTM6 in PDAC according to cBioPortal
   ([86]http://www.cbioportal.org/); B) Bioinformatics analysis using
   Cistrome Data Browser ([87]http://cistrome.org/) indicating enrichment
   of H3K27ac in the promoter of CMTM6 in CFPAC‐1 cells; C) H3K27ac
   ChIP‐seq data showing H3K27ac enrichment in the promoter of CMTM6 in
   BxPC‐3 cells; D) Comparison of EP300 and CREBBP expression levels in WT
   and BxPC‐3/GR and CFPAC‐1/GR cells ([88]GSE140077); E) Representative
   images of IHC staining for EP300 in WT and GR PDXs. Scale Bar, 100 µm;
   F) Upregulation of EP300 expression in PDAC tumor tissues compared to
   normal pancreatic tissues (TCGA and GTEx datasets); G) Correlation
   between CMTM6 and EP300 mRNA expression (TCGA dataset); H) qPCR assays
   showing that mRNA levels of CMTM6 in BxPC‐3/GR and CFPAC‐1/GR cells
   decrease in a dose‐dependent manner under treatment with the EP300
   inhibitor C646; I) Western blotting showing that protein expression of
   CMTM6 in BxPC‐3/GR and CFPAC‐1/GR cells decreases in a dose‐dependent
   manner under C646 treatment; J) Downregulation of CMTM6 mRNA levels in
   EP300‐knockout BxPC‐3/GR and CFPAC‐1/GR cells; K) Western blotting
   showing that protein expression of H3K27ac and CMTM6 in EP300‐KD PDAC
   cells (GR). L) ChIP‐qPCR analysis showing reduced H3K27ac enrichment in
   the CMTM6 promoter following EP300 knockdown in BxPC‐3/GR and
   CFPAC‐1/GR cells. Results represent three independent experiments in D
   and H‐L. Error bars represent the mean ± SD. ^** p < 0.01; ^*** p
   < 0.001 according to Student's t‐test.

   Using the UCSC genome bioinformatics site
   ([89]http://genome.ucsc.edu/), we evaluated the epigenetic modification
   of CMTM6, noting a significant enrichment of the H3K27ac signal in the
   promoter region of CMTM6 in CFPAC‐1 cells (Figure [90]2B). We further
   identified the H3K27ac modification sites in BxPC‐3 cells, another PDAC
   cell line, using chromatin immunoprecipitation sequencing (ChIP‐seq).
   The results confirmed significant H3K27ac modification in the promoter
   region of CMTM6 (Figure [91]2C). EP300 and CREBBP are major H3K27
   acetyltransferases that function as transcriptional activators.^[
   [92]^20 ^] However, we found that only EP300, but not CREBBP, is
   remarkably overexpressed in GEM‐resistant PDAC cells (Figure [93]2D).
   Similarly, in IHC assays, EP300 was more intensively stained in
   GEM‐resistant PDX tissues than in WT PDX tissues (Figure [94]2E).
   Additionally, analysis of TCGA datasets revealed that EP300 expression
   is upregulated in PDAC (Figure [95]2F) and is positively correlated
   with CMTM6 expression (Figure [96]2G). After treatment of C646, a
   histone acetyltransferase inhibitor targeting EP300, GEM‐resistant PDAC
   showed a dose‐dependent reduction in total H3K27ac and CMTM6 expression
   (Figure [97]2H,I). Similarly, following the selective knockdown of
   EP300, there was a significant decrease in the mRNA and protein levels
   of CMTM6 (Figure [98]2J,K) and in the enrichment of H3K27ac signals in
   the promoter region of CMTM6, as evidenced by ChIP assays
   (Figure [99]2L). These results indicate that EP300‐mediated epigenetic
   modification contributes to the transcriptional activation of CMTM6 in
   GEM‐resistant PDAC.

2.3. CMTM6 Promotes GEM Resistance in PDAC Both In Vitro and In Vivo

   To investigate the effect of CMTM6 on GEM resistance in PDAC, we
   knocked out and overexpressed CMTM6 in GEM‐resistant PDAC cells
   (Figure [100] 3A). Half‐maximal inhibitory concentration (IC[50])
   values indicated that CMTM6 knockdown restored the sensitivity to GEM,
   whereas its overexpression further promoted GEM resistance
   (Figure [101]3B; Figure [102]S2A, Supporting Information). Further
   results substantiated that CMTM6 enhanced cell proliferation
   (Figure [103]3C; Figure [104]S2B, Supporting Information), inhibited
   apoptosis (Figure [105]3D; Figure [106]S2C,D, Supporting Information),
   and promoted the clonogenic ability of GEM‐resistant PDAC cells
   following exposure to GEM (Figure [107]3E; Figure [108]S2E,F,
   Supporting Information).

Figure 3.

   Figure 3
   [109]Open in a new tab

   CMTM6 promotes GEM resistance in PDAC both in vitro and in vivo. A)
   qPCR and Western blotting verification of CMTM6 knockout and
   overexpression efficacy in BxPC‐3/GR and CFPAC‐1/GR cells; B) Influence
   of CMTM6 knockout and overexpression on IC[50] values of GEM in
   BxPC‐3/GR cells; C‐E. Effects of CMTM6 knockout and overexpression on
   cell growth rate (C), apoptosis rates (D) and colony‐formation
   abilities (E) in BxPC‐3/GR cells treated with 500 nm GEM; F) Mice
   harboring xenografts derived from CMTM6 knockout and overexpressing
   BxPC‐3/GR cells were treated with GEM (25 mg kg^−1, twice a week,
   intraperitoneally). Images of dissected tumors (left). Tumor growth
   curves, tumor weights, and mouse weights across groups (right); G)
   Representative images of IHC staining for Ki67 in xenografts from each
   group. Scale Bars, 200 µm (top) and 50 µm (bottom); H) Representative
   images of TUNEL analysis in xenografts from each group. Scale Bars,
   60 µm. Results represent three independent experiments in A–E, and
   results represent five samples in F–H. Error bars represent the mean ±
   SD. ^* p < 0.05; ^** p < 0.01; ^*** p < 0.001 according to Student's
   t‐test.

   Then, we performed in vivo studies, where mice harboring the indicated
   GEM‐resistant xenografts were treated with GEM (25 mg kg^−1,
   intraperitoneally, twice a week). Those harboring CMTM6‐knockout cells
   developed the smallest tumor, whereas those harboring
   CMTM6‐overexpressing cells developed the largest tumors
   (Figure [110]3F; Figure [111]S3A, Supporting Information). No
   significant difference in mouse weight was observed among the four
   groups. IHC staining of the tumor slices indicated that high expression
   of CMTM6 increased the proliferation (Figure [112]3G; Figure
   [113]S3B,C, Supporting Information) of GEM‐resistant PDAC cells in vivo
   and inhibited their apoptosis (Figure [114]3H; Figure [115]S3D,E,
   Supporting Information). Taken together, the above results demonstrate
   that CMTM6 confers GEM resistance in PDAC cells both in vitro and in
   vivo.

2.4. CMTM6 Protects IGF2BP1 from Ubiquitin‐Mediated Degradation

   The Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment
   analysis revealed a significant association between CMTM6 and the
   ubiquitin‐mediated proteolysis pathway in PDAC (Figure [116] 4A). To
   investigate the potential mechanism of CMTM6‐mediated resistance to
   GEM, we transfected GEM‐resistant BxPC‐3 cells with Flag‐tagged CMTM6
   and purified CMTM6‐bound protein complexes using anti‐Flag magnetic
   beads. Proteins associated with CMTM6 were identified using liquid
   chromatography‐tandem mass spectrometry (LC‐MS/MS) (Figure [117]4B).
   IGF2BP1, an m^6A reader, was detected in the purified CMTM6 complexes
   (Figure [118]4C). Co‐immunoprecipitation (Co‐IP) assays confirmed the
   interaction between endogenous CMTM6 and IGF2BP1 (Figure [119]4D;
   Figure [120]S4A, Supporting Information). Immunofluorescence analysis
   confirmed the co‐localization of CMTM6 and IGF2BP1 in GEM‐resistant
   PDAC cells (Figure [121]4E). Western blotting and quantitative
   polymerase chain reaction (qPCR) assays demonstrated that CMTM6
   increases IGF2BP1 protein levels without changing its mRNA expression
   (Figure [122]4F; Figure [123]S4B,C, Supporting Information). TCGA data
   also showed no correlation between the mRNA levels of CMTM6 and IGF2BP1
   (Figure [124]S4D, Supporting Information). These findings suggest that
   CMTM6 directly binds to IGF2BP1 and regulates its protein expression
   post‐transcriptionally.

Figure 4.

   Figure 4
   [125]Open in a new tab

   CMTM6 protects IGF2BP1 from ubiquitin‐mediated degradation. A) KEGG
   pathway enrichment analysis showing the pathways activated by CMTM6; B)
   CMTM6‐bound proteins in BxPC‐3/GR cells were purified by Co‐IP and
   identified by LC‐MS/MS. Scatterplots display proteins identified by
   LC‐MS/MS; C) The best unique peptide‐spectrum matches (PSM) of IGF2BP1;
   D) Co‐IP assays confirming the combination of CMTM6 and IGF2BP1 in
   BxPC‐3/GR cells; E) IF staining revealing the co‐localization of CMTM6
   and IGF2BP1 in BxPC‐3/GR cells; F) Western Blotting showing the
   expression levels of IGF2BP1 in CMTM6 knockout and overexpressing
   BxPC‐3/GR cells; G) CMTM6 knockout and overexpressing BxPC‐3/GR cells
   were treated with 10 µm CHX to block protein synthesis, and the
   degradation rate of IGF2BP1 was measured by Western Blotting (left).
   The half‐life of IGF2BP1 protein is shown (right); H) CMTM6 knockout
   and overexpressing BxPC‐3/GR cells were treated with DMSO or 20 µm
   MG132 for 6 h to block proteasome‐mediated protein degradation, and the
   synthesis rate of IGF2BP1 was measured by Western Blotting; I) CMTM6
   knockout and overexpressing BxPC‐3/GR cells were treated with 20 µm
   MG132 for 6 h, and the ubiquitination levels of IGF2BP1 were measured
   by IP assays. The results represent three independent experiments.

   To determine whether CMTM6 affects the stability of IGF2BP1, we treated
   GEM‐resistant cells with cycloheximide (CHX) to block protein
   biosynthesis and analyzed the degradation rate of IGF2BP1 at various
   CMTM6 expression levels. CMTM6 extended the half‐life of IGF2BP1
   (Figure [126]4G; Figure [127]S4E, Supporting Information), and the
   proteasome inhibitor MG132 reversed these changes in IGF2BP1 levels
   (Figure [128]4H; Figure [129]S4F, Supporting Information).

   Based on the KEGG analysis, we hypothesized that CMTM6 inhibits the
   ubiquitination of IGF2BP1, thereby stabilizing it. To test this, we
   used MG132 to block the proteasome‐mediated degradation of IGF2BP1 and
   examined its ubiquitination levels by immunoprecipitation.
   Overexpression of CMTM6 significantly reduced IGF2BP1 ubiquitination,
   whereas its knockdown increased IGF2BP1 polyubiquitination
   (Figure [130]4I; Figure [131]S4G, Supporting Information). These
   findings preliminarily confirm that CMTM6 inhibits the degradation of
   IGF2BP1 in PDAC cells by reducing its polyubiquitination.

2.5. IGF2BP1 Is Responsible for CMTM6‐Mediated GEM Resistance in PDAC

   To test the effect of IGF2BP1 on GEM resistance, we first analyzed data
   from TCGA and found that IGF2BP1 is overexpressed in PDAC (Figure [132]
   5A). IHC staining confirmed that an increase in IGF2BP1 levels in
   GEM‐resistant PDX (Figure [133]5B). Subsequently, we established
   IGF2BP1 knockout and overexpressing cell lines in GEM‐resistant PDAC
   cells (Figure [134]5C) and found that IGF2BP1 promoted cell
   proliferation following GEM exposure in vitro (Figure [135]5D).

Figure 5.

   Figure 5
   [136]Open in a new tab

   IGF2BP1 is responsible for CMTM6‐mediated GEM resistance in PDAC. A)
   Upregulation of IGF2BP1 expression in PDAC tumor tissues compared to
   normal pancreatic tissues (TCGA and GTEx datasets); B) Representative
   images of IHC staining for IGF2BP1 in WT and GR PDXs, Scale Bars,
   100 µm; C) qPCR verification of IGF2BP1 knockout and overexpression
   efficacy in BxPC‐3/GR and CFPAC‐1/GR cells; D) Influence of IGF2BP1
   knockout and overexpression on the cell growth rate of BxPC‐3/GR and
   CFPAC‐1/GR cells treated with 500 nm GEM; E) IGFB2P1 was overexpressed
   in CMTM6knockout cells. Cell growth curve showing that the cell
   proliferation of BxPC‐3/GR and CFPAC‐1/GR cells treated with 500 nm GEM
   is inhibited by CMTM6 knockout and rescued by IGF2BP1 overexpression;
   F) CMTM6 knockout promotes cell apoptosis in BxPC‐3/GR cells treated
   with 500 nm GEM, which is rescued by IGF2BP1 overexpression; G) IGF2BP1
   was silenced using shRNA or BTYNB in CMTM6‐overexpressing cells. Cell
   growth curve showing that proliferation of BxPC‐3/GR and CFPAC‐1/GR
   cells treated with 500 nm GEM is accelerated by CMTM6 overexpression,
   which is abolished by IGF2BP1 silencing; H) CMTM6 overexpression
   represses cell apoptosis in BxPC‐3/GR cells treated with 500 nm GEM,
   which is rescued by IGF2BP1 silencing. The results represent three
   independent experiments. Error bars represent the mean ± SD. ^** p
   < 0.01; ^*** p < 0.001 according to Student's t‐test.

   To further determine whether IGF2BP1 is responsible for the
   CMTM6‐mediated GEM resistance in PDAC cells, we overexpressed IGF2BP1
   in CMTM6‐knockout cells. Knockdown of CMTM6 significantly inhibited
   cell proliferation and promoted apoptosis, and these effects were
   rescued by IGF2BP1 overexpression (Figure [137]5E,F; Figure [138]S5A,
   Supporting Information). Additionally, we silenced IGF2BP1 using shRNA
   or BTYNB, a specific IGF2BP1 inhibitor, in CMTM6‐overexpressing cells.
   Similar to the above results, the accelerated cell proliferation and
   reduced cell apoptosis induced by CMTM6 overexpression were abolished
   by IGF2BP1 silencing (Figure [139]5G,H; Figure [140]S5B, Supporting
   Information). Collectively, these results indicate that IGF2BP1 is
   responsible for CMTM6‐mediated GEM resistance in PDAC.

2.6. IGF2BP1 Maintains EP300 and MYC mRNA Expression via m^6A Modification

   To identify the downstream targets of IGF2BP1, we performed methylated
   RNA immunoprecipitation‐sequencing (MeRIP‐seq) to identify mRNAs with
   m^6A modification and RNA immunoprecipitation‐sequencing (RIP‐seq) with
   an anti‐IGF2BP1 antibody to identify mRNAs binding to IGF2BP1 in
   GEM‐resistant PDAC cells. Additionally, RNA sequencing was conducted on
   IGF2BP1‐knockout PDAC cells to identify the mRNAs downregulated
   following IGF2BP1 knockdown.

   As shown in the Venn diagram, 92 genes, including EP300 and MYC, were
   screened as potential targets of IGF2BP1 (Figure [141] 6A). Integrative
   genomics viewer plots show the m^6A modification sites and IGF2BP1
   binding sites in GEM‐resistant PDAC cells (Figure [142]6B; Figure
   [143]S6A, Supporting Information). qPCR assays confirmed that the mRNA
   levels of MYC and EP300 were upregulated by IGF2BP1 (Figure [144]6C;
   Figure [145]S6B, Supporting Information).

Figure 6.

   Figure 6
   [146]Open in a new tab

   IGF2BP1 maintains EP300 and MYC mRNA expression via m^6A Modification.
   A) Venn diagram illustrating the overlap of genes containing m^6A and
   IGF2BP1 binding sites, and those downregulated by IGF2BP1 knockout in
   PDAC cells; B) Integrative genomics viewer (IGV) plots visualizing the
   m^6A binding sites and IGF2BP1 binding sites on EP300 and MYC mRNAs in
   BxPC‐3/GR cells; C) Quantification of EP300 and MYC mRNA expression
   levels after IGF2BP1 knockout and overexpression in BxPC‐3/GR cells; D)
   IGF2BP1‐knockout BxPC‐3/GR cells were treated with 5 µg mL^−1
   actinomycin D (ATCD) to block RNA synthesis, and the decay rate of
   EP300 and MYC mRNAs was measured by qPCR; E) RIP‐qPCR analysis using an
   IGF2PB1‐specific antibody and IgG control antibody confirming that
   IGF2BP1 binds to the mRNAs of EP300 and MYC; F) meRIP‐qPCR analysis
   using an m^6A‐specific antibody confirming m^6A binding sites on EP300
   mRNA; G) Influence of CMTM6 knockout on the sphere formation efficiency
   of BxPC‐3/GR and CFPAC‐1/GR cells treated with 500 nm GEM; H) mIF
   staining for CMTM6 (red), IGF2BP1 (green) and EP300 (yellow) in
   GEM‐resistant and GEM‐sensitive needle biopsy specimens obtained before
   chemotherapy (n = 20). Scale Bars, 100 µm. The results represent three
   independent experiments. Error bars represent the mean ± SD. ^* p
   < 0.05; ^** p < 0.01; ^*** p < 0.001 according to Student's t‐test.

   Then, we used actinomycin D (ACTD) to block mRNA biosynthesis and found
   that IGF2BP1 knockdown significantly reduced the stability of EP300 and
   MYC mRNAs (Figure [147]6D; Figure [148]S6C, Supporting Information).
   RIP‐qPCR assays showed that in GEM‐resistant PDAC cells, EP300 and MYC
   mRNAs were effectively precipitated by the anti‐IGF2BP1 antibody, and
   IGF2BP1 knockdown significantly decreased their enrichment
   (Figure [149]6E; Figure [150]S6D, Supporting Information). MeRIP‐qPCR
   assays using an m6A‐specific antibody confirmed the existence of m^6A
   modification sites on the mRNA of EP300 (Figure [151]6F).

   It is well known that MYC plays a crucial role in promoting stemness,
   which is associated with chemoresistance.^[ [152]^21 ^] To further
   investigate the mechanism underlying gemcitabine resistance, we
   analyzed MYC expression levels in PDAC cells with CMTM6 knockdown or
   overexpression. Our results demonstrated that CMTM6 knockdown led to a
   reduction in MYC RNA levels, while overexpression of CMTM6 elevated MYC
   RNA levels (Figure [153]S7A, Supporting Information). Moreover, CMTM6
   knockdown inhibited the sphere‐formation ability of PDAC cells exposed
   to GEM (Figure [154]6G; Figure [155]S7B, Supporting Information). Gene
   Set Enrichment Analysis (GSEA) using the TCGA dataset indicated a
   positive correlation between CMTM6 expression and the MYC TARGETS
   pathway (Figure [156]S7C, Supporting Information). We further examined
   the relationship between CMTM6 and MYC downstream target genes.
   Correlation analysis revealed that CMTM6 expression is significantly
   associated with PROM1 (CD133) and BMI1, but not with NANOG, SOX2, or
   POU5F1 (OCT4) (Figure [157]S7D, Supporting Information). PCR results
   confirmed that CMTM6 knockdown resulted in a marked downregulation of
   stemness‐associated genes, including BMI1 and PROM1 (CD133), while
   CMTM6 overexpression promoted the transcription of these genes (Figure
   [158]S7E, Supporting Information). Additionally, our data indicated
   that MYC depletion did not affect CMTM6 expression, suggesting that MYC
   does not regulate CMTM6 expression (Figure [159]S8, Supporting
   Information). In summary, our findings suggest that CMTM6
   overexpression enhances the expression of stemness‐related genes such
   as BMI1 and PROM1 in PDAC cells, thereby conferring increased
   resistance to chemotherapy.

   Furthermore, we obtained needle aspiration samples from PDAC patients
   prior to treatment with GEM‐based chemotherapy. Among these patients, 4
   showed sensitivity to GEM‐based chemotherapy in subsequent treatment,
   whereas 16 exhibited drug resistance. mIF analysis showed that CMTM6,
   EP300, and IGF2BP1 were highly expressed in chemo‐resistant PDAC
   samples (Figure [160]6H; Figure [161]S9A,B, Supporting Information). A
   positive correlation was observed among the levels of the three
   proteins (Figure [162]S9C, Supporting Information). In particular, we
   noted that EP300 is both an upstream protein activating CMTM6
   transcription and a downstream target gene of IGF2BP1. Our results
   suggest that IGF2BP1 directly binds to the mRNAs of EP300 and MYC,
   maintaining their stability via an m^6A‐dependent manner. This forms a
   positive feedback loop of EP300–CMTM6–IGF2BP1–EP300 (mRNA), which
   enhances tumor stemness and ultimately facilitates PDAC resistance.

2.7. Blocking the CMTM6/EP300 Positive Feedback Loop Sensitizes PDAC Cells to
GEM

   Given the pivotal role of EP300‐mediated epigenetic modification in the
   EP300–CMTM6–IGF2BP1–EP300 (mRNA) positive feedback loop, we
   hypothesized that targeting EP300 might be a novel therapeutic approach
   in GEM‐resistant PDAC. Inobrodib, the most recent inhibitor developed
   to target EP300, is the first to undergo phase I/IIa clinical trials
   for advanced solid tumors.^[ [163]^22 ^] In this study, we found that
   inobrodib effectively suppressed the survival of PDAC cells, even at a
   relatively low dose (Figure [164]S10A, Supporting Information).

   To evaluate the feasibility of the combination treatment of inobrodib
   and GEM, we determined the combination index (CI) of the two drugs in
   GEM‐resistant cells. Inobrodib and GEM exhibited a synergistic
   therapeutic effect (Figure [165] 7A,B). Subsequent cell viability
   (Figure [166]7C; Figure [167]S10B, Supporting Information) and
   apoptosis (Figure [168]7D; Figure [169]S10C,D, Supporting Information)
   assays indicated that their combined use had stronger effects on cell
   proliferation inhibition and apoptosis promotion than monotherapy
   alone.

Figure 7.

   Figure 7
   [170]Open in a new tab

   Blocking the CMTM6/EP300 positive feedback loop sensitizes PDAC cells
   to GEM. A) Combination Index (CI) depicting the synergistic effect of
   GEM and Inobrodib on BxPC‐3/GR cells; B) Dose‐response curves
   illustrating the individual and combined treatment effects of GEM and
   Inobrodib on BxPC‐3/GR cells; C) Cell growth curves of BxPC‐3/GR cells
   treated with PBS, Inobrodib, GEM, or the combination of GEM and
   Inobrodib; D) Apoptosis assays revealing the effects of PBS, Inobrodib,
   GEM, and the combination of GEM and Inobrodib on BxPC‐3/GR cells; E)
   Mice harboring WT PDXs were treated with PBS, Inobrodib, GEM, and the
   combination of GEM and Inobrodib. Tumor photos, tumor weights, tumor
   volume growth curves, and mouse weights across groups are displayed. F)
   Mice harboring GR PDXs were treated with PBS, Inobrodib, GEM, and the
   combination of GEM and Inobrodib. Tumor photos, tumor weights, tumor
   volume growth curves, and mouse weights across groups are presented. G)
   Statistical analysis showing blood routine examination, liver function,
   and kidney function of mice in different treatment groups; H) Typical
   images of IHC staining for Ki67 in GR PDXs from different treatment
   groups. Scale Bars, 200 µm (top) and 50 µm (bottom); I) Representative
   images of TUNEL analysis in GR PDXs from different treatment groups.
   Scale Bars, 60 µm. Results represent three independent experiments in
   A‐D, and results represent five samples in D‐I. Error bars represent
   the mean ± SD. ^** p < 0.01; ^*** p < 0.001; ns means p > 0.05
   according to Student's t‐test.

   Furthermore, we used WT and GEM‐resistant PDX to assess this
   synergistic therapeutic effect in vivo. As shown in Figure [171]7E,F,
   mice treated with inobrodib plus GEM developed the smallest tumors in
   both groups. Compared to GEM monotherapy, the combined treatment had no
   significant effects on mouse weights, and routine blood, liver, and
   kidney function tests (Figure [172]7G). IHC assays revealed the weakest
   staining of the proliferation marker gene Ki67 in the combined
   treatment group (Figure [173]7H; Figure [174]S10E,F, Supporting
   Information), while the TUNEL assays showed the highest apoptotic rates
   in this group (Figure [175]7I; Figure [176]S10G,H, Supporting
   Information). These results imply that the EP300 inhibitor inobrodib
   may synergistically improve the antitumor effect of GEM without
   increasing side effects.

3. Discussion

   Owing to its low resection rate and high recurrence rate, chemotherapy
   plays a crucial role in the management of PDAC.^[ [177]^23 ^] GEM is
   the current standard chemotherapy agent for advanced PDAC^[ [178]^24
   ^]; however, the development of GEM resistance limits its efficacy in
   patients. GEM is a genotoxic agent that interferes with cellular DNA
   synthesis and blocks the cell cycle.^[ [179]^25 ^] Nevertheless, the
   mechanism underlying GEM resistance remains unclear. Therefore, there
   is an urgent need to elucidate the molecular mechanisms of GEM
   resistance and identify new therapeutic strategies in PDAC. The PDX
   model preserves the heterogeneity of primary tumors and the composition
   of the tumor microenvironment, making it the most appropriate model for
   studying tumor chemoresistance. In this study, we established
   GEM‐resistant PDX models of PDAC and GEM‐resistant PDAC cells. Analysis
   of sequencing results, along with TCGA and GEO datasets, revealed that
   CMTM6 is overexpressed in GEM‐resistant PDAC, and this was confirmed in
   PDAC clinical specimens and PDX models.

   Histone modification, a common epigenetic alteration of chromatin DNA,
   not only directly regulates transcription but also affects other
   processes, such as DNA repair, DNA replication, stemness, and changes
   in cell state.^[ [180]^15 ^] While exploring the mechanism behind CMTM6
   upregulation in GEM‐resistant PDAC using bioinformatics analysis and
   ChIP‐seq, we found a significant enrichment of the H3K27ac signal in
   the promoter region of CMTM6. This acetylation of lysine is highly
   dynamic and controlled by the opposing actions of histone
   acetyltransferases and histone deacetylases.^[ [181]^26 ^] EP300
   (KAT3B) and CREBBP (KAT3A) are the two major acetyltransferases
   responsible for histone acetylation. This study revealed that
   EP300‐mediated H3K27ac modification significantly enhances CMTM6
   expression. The expression of CMTM6 increased as the GEM treatment was
   prolonged, resulting in acquired chemoresistance in PDAC. Thus, this
   study provides new evidence on how epigenetic modification mediates
   acquired resistance in tumors.

   CMTM6 can maintain PD‐L1 expression by preventing its
   ubiquitination‐mediated degradation.^[ [182]^6 ^] In this study, we
   found that CMTM6 promoted GEM resistance of PDAC both in vitro and in
   vivo. Furthermore, there was a strong correlation between CMTM6 and the
   ubiquitination–proteasome pathway in PDAC. Therefore, we performed
   Co‐IP assays, followed by LC‐MS/MS, to identify the proteins
   interacting with CMTM6. Intriguingly, CMTM6 could directly bind to
   IGF2BP1 and extend its half‐life by protecting IGF2BP1 from
   ubiquitination.

   Numerous studies have revealed the important role of m^6A modification
   in tumor malignancy and drug resistance.^[ [183]^27 ^] However, the
   post‐transcriptional modification of m^6A‐related proteins remains
   controversial. Sun et al. found that phosphorylation of METTL3 enhanced
   its stability and activity, whereas ubiquitination promoted its
   degradation.^[ [184]^28 ^] Ubiquitination also mediates the degradation
   of METTL14^[ [185]^29 ^] and IGF2BP2.^[ [186]^30 ^] This study revealed
   that IGF2BP1 is degraded through the ubiquitination–proteasome pathway,
   which could be inhibited by CMTM6.

   As an m^6A reader, IGF2BP1 enhances RNA stability by recognizing m^6A
   modifications.^[ [187]^31 ^] IGF2BP1 is highly expressed and induces
   drug resistance in colorectal cancer,^[ [188]^32 ^] ovarian cancer,^[
   [189]^33 ^] and osteosarcoma^[ [190]^34 ^] and contributes to
   oxaliplatin resistance by enhancing stemness in gastric cancer.^[
   [191]^35 ^] Herein, we found that IGF2BP1 expression is upregulated in
   PDAC and positively correlates with GEM resistance. Furthermore,
   IGF2BP1 was responsible for CMTM6‐mediated GEM resistance.

   Tumor heterogeneity and the development of drug resistance have been
   attributed to the properties of cancer cell stemness, which pertains to
   the capacity to self‐renew, regenerate tumor heterogeneity, and
   increase the expression of drug‐resistance genes.^[ [192]^36 ^] MYC is
   an oncogene that regulates cell growth, differentiation, and
   chemoresistance by enhancing tumor cell stemness.^[ [193]^21 ^] It
   activates the transcription of stemness factors by binding to DNA
   recognition sequences in the promoter regions of target genes.^[
   [194]^37 ^] Moreover, MYC is a target of IGF2BP1.^[ [195]^19 , [196]^38
   ^] In this study, we confirmed that IGF2BP1 stabilizes MYC mRNA by
   recognizing its m^6A modifications. In this study, we confirmed that
   IGF2BP1 stabilizes MYC mRNA by recognizing its m^6A modifications.
   Furthermore, CMTM6 overexpression enhances the expression of
   stemness‐related genes, such as BMI1 and PROM1, in PDAC cells through
   MYC‐mediated transcriptional activation, thereby conferring increased
   resistance to gemcitabine.

   To our knowledge, this study is the first to reveal that the
   EP300‐mediated epigenetic modification significantly increased CMTM6
   expression, which repressed the ubiquitination of the IGF2BP1 protein.
   Moreover, IGF2BP1 stabilized EP300 mRNA in an m^6A‐dependent manner,
   forming a positive feedback loop comprising EP300–CMTM6–IGF2BP1–EP300
   (mRNA), thereby further enhancing tumor stemness and GEM resistance in
   PDAC. Our study described a novel axis that integrates histone
   modification of DNA, m^6A modification of mRNA, and ubiquitination of
   proteins. We explored the influence of modifications in these three
   dimensions on GEM resistance.

   We believe that the effect of CMTM6 on the development of GEM
   resistance is attributed to the amplification effect of the positive
   feedback loop. EP300 plays a central role in this loop, making it a
   potential therapeutic target for GEM‐resistant PDAC. EP300 inhibitors
   have emerged as promising novel antitumor agents for clinical
   translation.^[ [197]^22 ^] Specifically, inobrodib (CCS1477), an EP300
   bromodomain inhibitor, is under clinical evaluation for patients with
   advanced and drug‐resistant solid tumors or hematological malignancies.
   In this study, the efficacy and safety of the combination treatment of
   inobrodib and GEM were evaluated using a combination index in vitro and
   GEM‐resistant PDX models in vivo. The findings indicate that blocking
   the EP300–CMTM6 axis may represent a potential therapy for
   GEM‐resistant PDAC.

   In summary, the EP300‐mediated H3K27ac modification activates the
   transcription of CMTM6 in GEM‐resistant PDAC. Overexpressed CMTM6
   prevents the ubiquitination of the m^6A reader IGF2BP1, maintaining the
   stability of EP300 and MYC mRNAs in an m^6A‐dependent manner. The
   EP300–CMTM6–IGF2BP1 axis forms a positive feedback loop to amplify the
   effect of MYC in tumor stemness, ultimately facilitating PDAC
   resistance. Moreover, combined treatment of GEM and inobrodib is a
   potential strategy for GEM‐resistant PDAC. This study revealed a novel
   epigenetic mechanism behind chemoresistance in PDAC and suggests new
   therapeutic approaches with direct potential for clinical trials.

4. Experimental Section

Patients and Specimens

   All human samples were obtained from the First Affiliated Hospital of
   Sun Yat‐sen University (FAHSYSU) in China from January 2016 to December
   2020. As of the last follow‐up date of October 23, 2023, the follow‐up
   rate was 100%. A total of 45 paraffin‐embedded tissues from PDAC
   patients were collected before receiving GEM‐based chemotherapy. The
   inclusion criteria included: a pathological diagnosis of PDAC; no prior
   radiotherapy; and GEM‐based chemotherapy. According to the Response
   Evaluation Criteria in Solid Tumors (RECIST1.1),^[ [198]^39 ^] 16 of
   the 45 patients were evaluated as GEM‐resistant (progressive disease,
   PD), and 4 were evaluated as GEM‐sensitive (partial response, PR). The
   remaining 25 patients were categorized as having stable disease (SD).
   From November 2018 to July 2019, 8 consecutive specimens of fresh PDAC
   tissues were prospectively collected during surgery. The PDAC tissue
   mass was used to establish PDX models.

Establishment of GEM‐Resistant Cell Lines

   The 293T, BxPC‐3, and CFPAC‐1 cell lines were purchased from the Cell
   Bank of the Chinese Academy of Science (Shanghai, China). GEM was
   purchased from Topscience (T6069, Shanghai, China). WT PDAC cells,
   namely BxPC‐3 and CFPAC‐1, were inoculated into 96‐well plates and
   exposed to GEM in gradually increasing concentrations (ranging from
   10 nm to 1 µm) for 48 h. The level of cell viability was determined
   using the CellTiter‐Glo Luminescent Cell Viability Assay kit, and the
   IC[50] value of GEM was calculated for WT PDAC cell lines. The
   reasonable drug treatment concentration was determined according to the
   calculated IC[50] value. BxPC‐3 and CFPAC‐1 cells in the logarithmic
   growth phase were cultured in the medium containing 1 µm GEM for 1 h,
   and the medium was discarded. Cells were washed twice with PBS and
   replaced with a drug‐free medium for continued routine culture until
   the cells recovered logarithmic growth. These methods were repeated,
   and the exposure time to GEM was gradually prolonged (1, 2, 4, 6, 12,
   24, 36, 48, 60, and 72 h) for 6 months. Finally, the cells were able to
   grow and propagate stably in a medium containing 1 µm GEM. IC[50]
   assays confirmed the successful establishment of GEM‐resistant cell
   lines.

Establishment of GEM‐Resistant PDX

   Surgical specimens from PDAC patients were cut into 3 mm^3 pieces and
   engrafted into 5‐week‐old B‐NDG mice (Biocytogen, Beijing, China),
   designated as generation 1 (F1). Tumor volumes were measured using a
   vernier caliper and calculated as 0.5 × length × width^2. Mice were
   euthanized when the transplanted tumor volume reached 1 cm^3. The
   subcutaneous transplanted tumors were collected, cut into smaller
   pieces, and transplanted into a second batch of mice for expansion
   (F2). After tumor formation in F2, the mice were randomly divided into
   a PBS treatment group and a GEM treatment group (25 mg kg^−1,
   intraperitoneally, twice a week; F3).

Animal Experiments

   PDAC cells with either knockdown or overexpression of CMTM6 were
   subcutaneously implanted into 4‐week‐old BALB/c nude mice (Laboratory
   Animal Center of Sun Yat‐Sen University). When the tumor volumes
   reached 50 mm^3, the mice were treated with GEM intraperitoneally twice
   a week. Tumor volumes were recorded twice weekly, and mouse weights
   were recorded once weekly. At the end of the experiment (1 month after
   dosing), all mice were euthanized, and the tumors were excised,
   weighed, and embedded in paraffin. Serial 4.0‐mm sections were obtained
   and analyzed by immunohistochemistry (IHC) with a Ki67‐specific
   antibody and TdT‐mediated dUTP Nick‐End Labeling (TUNEL) assays.

   For the in vivo drug combination experiment, WT or GEM‐resistant PDX
   tumors were implanted into mice. When the tumor size reached 25 mm^3,
   mice were treated and randomized into four groups: PBS, inobrodib
   (30 mg kg^−1, orally, every other day), GEM (25 mg kg^−1,
   intraperitoneally, twice a week), and combined inobrodib and GEM
   treatment groups. Tumor volumes and mouse weights were recorded twice
   and once weekly, respectively. At the experimental endpoint (21 days
   after dosing), blood counts, and kidney and liver function tests were
   performed, and cell proliferation and apoptosis were validated via IHC
   and TUNEL assays, respectively.

Statistical Analysis

   Statistical analysis was performed in GraphPad Prism 8.0 and CompuSyn
   software. Data are expressed as the mean ± standard deviation (SD). The
   differences between the groups were evaluated using the Student
   two‐tailed t‐test and one‐way analysis of variance (ANOVA). The
   Spearman correlation coefficient was used to evaluate the correlation
   between the two groups. Statistical significance was set at p < 0.05.
   The effects of drug combinations were analyzed based on the combination
   index (CI). Significance levels are indicated as follows: ^* p < 0.05;
   ^** p < 0.01; ^*** p < 0.001; ns means p > 0.05.

Ethics Approval Statement

   All experiments involving humans were carried out in accordance with
   the Code of Ethics of the World Medical Association (Declaration of
   Helsinki). All procedures were performed in compliance with relevant
   laws and institutional guidelines. Ethics approval for the use of human
   tumor specimens was provided by the Internal Review and the Ethics
   Committee of the FAHSYSU (IIT‐2021‐719). Written informed consent was
   obtained from all patients before sample collection.

   All animal experiments were approved by the Institutional Review Board
   of FAHSYSU (2021‐498). The animals received humane care according to
   the guidelines outlined in the “Guide for the Care and Use of
   Laboratory Animals”.

Conflict of Interest

   The authors declare no conflict of interest.

Author Contributions

   Y.‐Q.Z., Y.H., and Y.‐H.S. contributed equally to this work, X.‐Y.Y.
   conceived the study. X.‐Y.Y. Y.‐Q.Z., Y.H., Y.‐H.S., and Q.‐C.X.
   designed the study. Y.‐Q.Z., Y.H., Y.‐H.S., and G.‐Y.Z. performed
   experiments and analyzed the results. C.‐S.H., Z.‐D.L., M.‐J.M.,
   J.‐Y.Y., X.X., Q.L., X.‐T.H., and J.‐Q.W. assisted in the collection
   and analysis of clinical samples. X.‐Y.Y. and Q.‐C.X. supervised and
   guaranteed the study. X.‐Y.Y., Y.‐Q.Z., and Y.H. wrote the manuscript.
   All authors commented on the manuscript and approved the final version
   of the manuscript.

Supporting information

   Supporting Information
   [199]ADVS-11-2406714-s001.pdf^ (558.4KB, pdf)

Acknowledgements