Abstract

   Brain metastases represent a significant cause of morbidity and
   mortality in non-small cell lung cancer (NSCLC), with limited
   therapeutic options. The unique brain microenvironment, characterized
   by low lipid availability, may drive NSCLC cells to adapt through lipid
   metabolic reprogramming. In this study, we identify a novel mechanism
   by which IGF2BP3-driven lipid metabolism promotes the brain
   colonization of NSCLC cells through the IGF2BP3-FASN axis. Elevated
   IGF2BP3 expression in NSCLC brain metastases correlates with poor
   prognosis and promotes cancer cell migration, invasion, and brain
   colonization by activating the lipogenesis pathway. We further
   identified that FASN was a downstream target of IGF2BP3 in NSCLC cells.
   Mechanistically, IGF2BP3 binds to FASN mRNA, enhancing its stability
   through RNA-binding activity. FASN is essential for neutral lipid
   accumulation and brain colonization, as demonstrated in vitro and in
   vivo. Our findings highlight the critical role of IGF2BP3 in lipid
   metabolism and propose that targeting IGF2BP3 may provide a promising
   therapeutic strategy for NSCLC brain colonization.

   Subject terms: Non-small-cell lung cancer, Non-small-cell lung cancer

Introduction

   Lung cancer is the leading cause of cancer-related mortality worldwide,
   accounting for 18% of cancer deaths [[52]1]. Non-small cell lung cancer
   (NSCLC) constitutes 80–85% of all lung cancer cases [[53]2] and is
   frequently associated with brain metastases, occurring in 10–20% of
   patients at diagnosis and up to 40% during disease progression [[54]3].
   Current treatment options, including surgery, stereotactic
   radiotherapy, whole-brain radiotherapy, and chemotherapy, have limited
   efficacy against brain metastatic lesions, resulting in poor prognoses
   [[55]4, [56]5]. The colonization step is a crucial part of brain
   metastasis formation, which depends on the survival and proliferation
   abilities of NSCLC cells within the brain microenvironment. Thus,
   understanding the mechanisms of brain colonization of NSCLC cells is
   essential for developing effective therapeutic strategies.

   Metabolic reprogramming is a hallmark of cancer progression and
   metastasis [[57]6]. Adaptation to the brain’s unique nutrient-deprived
   microenvironment, largely due to the restrictive blood-brain barrier
   (BBB), requires metabolic adjustments for cancer cells to survive and
   grow [[58]7–[59]9]. The brain’s limited extracellular lipid
   availability necessitates de novo lipid synthesis, a process heavily
   reliant on fatty acid synthase (FASN) [[60]10]. FASN catalyzes the
   synthesis of fatty acids from acetyl-CoA, malonyl-CoA, and NADPH
   [[61]11]. Elevated FASN expression has been observed in various
   cancers, including lung, breast, colorectal, and cholangiocarcinoma
   [[62]12–[63]16]. Xu and colleagues observed that FASN were
   significantly enriched in cholangiocarcinoma and FASN inhibitor
   significantly reversed the immunosuppressive microenvironment and
   enhanced anti-PD-1 efficacy in mouse model [[64]13]. Recently, it has
   been reported that inhibition of FASN effectively reduce the brain
   metastasis burden in human epidermal growth factor receptor 2-positive
   breast cancer [[65]17]. However, its role in NSCLC brain
   microenvironment adaption remains unclear.

   RNA-binding proteins (RBPs) regulate post-transcriptional processes
   [[66]18] and contribute to cancer metabolism and metastasis [[67]19,
   [68]20]. Insulin-like growth factor 2 mRNA-binding protein 3 (IGF2BP3)
   is a member of RBPs consisting of two N-terminal RNA recognition
   binding regions (RRM) and four C-terminal heterologous
   ribonucleoprotein KH domains [[69]21, [70]22]. IGF2BP3 is associated to
   tumor growth and migration in many kinds of cancer and related to poor
   prognosis [[71]22–[72]29]. Meanwhile, IGF2BP3 has been reported to be
   involved in tumor metastasis progression in gastric cancer, bladder
   cancer, cervical cancer and hepatocellular cancer [[73]30–[74]33].
   Importantly, recent studies suggested that IGF2BP3 may be associated
   with the progression and prognosis of brain metastases in breast cancer
   and NSCLC [[75]34, [76]35]. Nevertheless, the specific mechanisms of
   IGF2BP3 regulating brain colonization remain elusive.

   In this study, we demonstrated that IGF2BP3-mediated lipid metabolism
   played a critical role in NSCLC brain colonization. IGF2BP3 is
   upregulated in NSCLC brain metastases and correlated with poor
   prognosis. Inhibition of IGF2BP3 triggers a remarkable reprogramming of
   cellular metabolites, especially those involved in the lipid
   metabolism, resulting in the reduction of total triglycerides,
   cholesterol, lipid droplets and neutral lipid level of NSCLC cells.
   Furthermore, we revealed that targeting IGF2BP3 significantly
   suppressed NSCLC cell growth within the brain microenvironment and
   neutral lipid accumulation in mouse model and this effect was dependent
   on the FASN, a key downstream target of IGF2BP3 in lipogenesis. These
   findings establish the IGF2BP3–FASN axis as a critical regulator of
   NSCLC lipid metabolism and brain colonization and highlight IGF2BP3 as
   a potential therapeutic target in NSCLC brain metastasis.

Results

IGF2BP3 is upregulated in NSCLC brain metastasis and associated with the poor
prognosis

   To investigate the role of IGF2BP3 in NSCLC brain metastasis, we
   analyzed multiple NSCLC cohorts with metastasis using data from the
   TCGA and GEO databases. IGF2BP3 expression was significantly
   upregulated in advanced metastatic stages in the TCGA cohort (Fig.
   [77]1A). Consistent with this, IGF2BP3 levels were higher in brain
   metastases compared to primary lung tumors in NSCLC from the GEO
   cohorts ([78]GSE110495 and [79]GSE83132) (Fig. [80]1B, C). Kaplan-Meier
   survival analysis further revealed that elevated IGF2BP3 expression
   correlated with poor prognosis in NSCLC patients with metastasis (Fig.
   [81]1D). Pearson correlation analysis demonstrated significant positive
   associations between IGF2BP3 and key epithelial-mesenchymal transition
   (EMT) markers, including TGFB1, SNAI2, N-cadherin, and MMP9, in human
   NSCLC tumors (Fig. [82]1E–H). To validate these findings in vivo, we
   established subcutaneous tumor model and a stereotactic brain injection
   model in mice and assessed IGF2BP3 protein expression (Fig. [83]1I).
   Immunohistochemistry (IHC) staining revealed significantly higher
   IGF2BP3 levels in brain colonization tumor compared to subcutaneous
   tumors (Fig. [84]1J, K). Concurrently, MMP9 protein expression was also
   elevated in brain colonization tumor (Fig. [85]1J, K). Collectively,
   these findings suggest a potential role for IGF2BP3 in regulating NSCLC
   brain metastasis.

Fig. 1. IGF2BP3 is upregulated in NSCLC brain metastasis and associated with
poor prognosis.

   [86]Fig. 1
   [87]Open in a new tab

   A IGF2BP3 expression was elevated in N2 (n = 74) and N3 (n = 2) stage
   metastatic tumor compared with N1 stage metastatic tumor (n = 95) in
   lung adenocarcinoma (LUAD) patients with lymph node metastasis
   according to TCGA database. Data are presented as median value. B
   IGF2BP3 expression levels were elevated in brain metastatic tumors (BT)
   (n = 6) compared with primary lung tumors (LT) (n = 8) in early passage
   cell lines derived from primary patient samples of lung-to-brain
   metastases according to [88]GSE110495 dataset. Data are presented as
   median value. C IGF2BP3 expression levels were elevated in LLC-derived
   brain metastatic cells (Brm) compared with primary cells according to
   [89]GSE83132 dataset. Data are presented as mean ± SD; n = 3. D Overall
   survival analysis of LUAD patients at the N1 stage based on TCGA
   project was calculated with the log-rank test and stratified by high or
   low IGF2BP3 expression (high: top 50%, n = 127; low: bottom 50%,
   n = 126). E–H Pearson correlation analysis between IGF2BP3 expression
   and key EMT markers (TGFB1, SNAI2, N-cadherin, and MMP9) in LUAD
   patients according to [90]GSE33532 dataset (n = 100). I Schematic of
   the subcutaneous tumor and stereotactic brain injection tumor model. J
   Hematoxylin and eosin (H&E) staining and immunohistochemistry (IHC)
   analysis of IGF2BP3 and MMP9 in representative subcutaneous tumor and
   xenograft tumor in the brain from mice. Scale bar = 20 μm. K
   Quantification of IGF2BP3 and MMP9 expression levels from IHC analysis
   in J using Image Pro Plus. Data are presented as mean ± SD; n = 3.
   *P < 0.05; **P < 0.01; ****P < 0.0001 (two-tailed t-test).

IGF2BP3 promotes the growth of NSCLC cells in the brain

   To determine the functional role of IGF2BP3 in NSCLC invasiveness, we
   generated IGF2BP3-knockdown (shIGF2BP3) and IGF2BP3-overexpressing
   (OE-IGF2BP3) H1299 and H157 cell lines. Knockdown and overexpression
   were confirmed at the mRNA and protein levels by RT-qPCR and
   immunoblotting, respectively (Figs. S[91]1A–D and S[92]3A–D).
   Functional assays demonstrated that IGF2BP3 knockdown reduced cell
   proliferation (Fig. S[93]1E–H) and colony formation (Fig. S[94]1I–K)
   while increasing apoptosis (Fig. S[95]1M–P). Importantly, IGF2BP3
   knockdown significantly inhibited cell migration (Figs. [96]2A, B and
   S[97]2A, B) and invasion (Fig. [98]2C, D and S[99]2C, D). Conversely,
   IGF2BP3 overexpression enhanced proliferation (Fig. S[100]3E–H), colony
   formation (Fig. S[101]3I–K), migration (Fig. [102]2E, F and S[103]2E,
   F) and invasion (Fig. [104]2G, H and S[105]2G, H), while reducing
   apoptosis (Fig. S[106]3L–O) in H1299 and H157 cell lines.

Fig. 2. IGF2BP3 promotes the growth of NSCLC cells in the brain.

   [107]Fig. 2
   [108]Open in a new tab

   A Wound healing assay showing cell migration ability in H1299 cells
   treated with shRNAs targeting IGF2BP3 or control (shCtrl) after 24 h.
   Scale bars = 50 μm. B Quantification of cell migration in
   IGF2BP3-knockdown H1299 cells from (A) using ImageJ. Data are presented
   as mean ± SD; n = 3. C Transwell invasion assay showing cell invasion
   ability in H1299 cells treated with shRNAs targeting IGF2BP3 or control
   after 24 h. Scale bars = 10 μm. D Quantification of cell invasion in
   IGF2BP3-knockdown H1299 cells from (C) using ImageJ. Data are presented
   as mean ± SD; n = 3. E Wound healing assay showing cell migration
   ability in H1299 cells overexpressing IGF2BP3 (OE-IGF2BP3) or control
   after 24 h. Scale bars = 50 μm. F Quantification of cell migration in
   IGF2BP3-overexpressing H1299 cells from (E) using ImageJ. Data are
   presented as mean ± SD; n = 3. G Transwell invasion assay showing cell
   invasion ability in H1299 cells overexpressing IGF2BP3 or control after
   24 h. Scale bars = 10= μm. H Quantification of cell invasion in
   IGF2BP3-overexpressing H1299 cells from (G) using ImageJ. Data are
   presented as mean ± SD; n = 3. I–K. H1299-luc cells (5 × 10^5) with
   IGF2BP3 knockout (sgIGF2BP3) or control (sgCtrl) were injected into
   BALB/c-nu/nu mice via stereotactic brain injection (n = 6 per group).
   Representative bioluminescent images of NSCLC xenograft tumor in the
   brain captured by the Neurostar IVIS preclinical in vivo imaging system
   on day 27 are shown in (I, J). Quantification of bioluminescence in the
   brain on day 27 is shown in (K). Data are presented as mean ± SD;
   n = 6. L–N H1299-luc cells (5 × 10^5) overexpressing IGF2BP3
   (OE-IGF2BP3) or control were injected into BALB/c-nu/nu mice via
   stereotactic brain injection (n = 6 per group). Representative
   bioluminescent images of NSCLC xenograft tumor in the brain on day 27
   are shown in (L, M). Quantification of bioluminescence in the brain on
   day 27 is shown in (N). Data are presented as mean ± SD; n = 6. O H&E
   staining and IHC analysis of IGF2BP3 and MMP9 in representative brain
   tumors from IGF2BP3-knockout and control mice. Scale bar = 20 μm. P
   Quantification of IGF2BP3 and MMP9 expression levels from IHC analysis
   in (O) using Image Pro Plus. Data are presented as mean ± SD; n = 3. Q
   IHC staining depicted the correlation between IGF2BP3 and MMP9
   expression in brain tumors from IGF2BP3-knockout and control mice
   (n = 3). R H&E staining and IHC analysis of IGF2BP3 and MMP9 in
   representative brain tumors from IGF2BP3-overexpressing and control
   mice. Scale bar = 20 μm. S Quantification of IGF2BP3 and MMP9
   expression levels from IHC analysis in (R) using Image Pro Plus. Data
   are presented as mean ± SD; n = 3. T IHC staining depicted the
   correlation between IGF2BP3 and MMP9 expression in brain tumors from
   IGF2BP3-overexpressing and control mice (n = 3). *P < 0.05; **P < 0.01;
   ***P < 0.001; ****P < 0.0001 (two-tailed t-test).

   Furthermore, we measured the ability of NSCLC cells to colonize the
   brain, and established a mouse stereotactic brain injection model using
   luciferase-tagged H1299 cells (Fig. S[109]2I). Firstly, we constructed
   stable IGF2BP3-knockout and IGF2BP3 overexpression H1299 cell lines
   with luciferase (sgIGF2BP3-luc and OE-IGF2BP3-luc, respectively). The
   protein expression of IGF2BP3 was verified by immunoblotting (Figs.
   S[110]1L and S[111]3P). H1299 cells were injected into the cortex of
   4-week-old nude mice and the body weight was measured every three days
   (Fig. S[112]2J, K). The tumors were imaged in vivo by IVIS system every
   nine days and plotted a growth curve. Next, the mice brains were
   harvested and imaged on day 27. IGF2BP3 knockout significantly reduced
   the growth of NSCLC cells in the brain (Figs. [113]2I–K, S[114]2L),
   whereas IGF2BP3 overexpression exacerbated tumor burden (Fig. [115]2L–N
   and S[116]2M). IHC staining confirmed IGF2BP3 expression changes in
   brain colonization tumor with MMP9 expression patterns aligning with
   IGF2BP3 levels (Fig. [117]2O–T). These results indicate that IGF2BP3
   promotes the growth of NSCLC cells in the brain, highlighting its
   potential as a therapeutic target.

IGF2BP3 mediates lipid metabolism in NSCLC

   To elucidate the mechanism by which IGF2BP3 promotes the growth of
   NSCLC cells in the brain, we performed transcriptome sequencing on
   IGF2BP3-knockdown (shIGF2BP3) and control (shCtrl) H1299 cells. We
   identified 160 upregulated and 469 downregulated genes (fold change
   ≥1.5, p ≤ 0.05) (Fig. [118]3A, B). Gene enrichment analysis revealed
   significant alterations in metabolic pathways (Fig. [119]3C). GSEA
   analysis further linked IGF2BP3 to fatty acid metabolism (Fig.
   [120]3D).

Fig. 3. IGF2BP3 mediates lipid metabolism in NSCLC.

   [121]Fig. 3
   [122]Open in a new tab

   A Volcano plot of differentially expressed genes (fold change > 1.5,
   p < 0.05) in H1299 cells stably expressing shRNAs targeting IGF2BP3
   (shIGF2BP3) or scramble control (shCtrl). B Heatmap of differentially
   expressed genes (fold change > 1.5, p < 0.05) between IGF2BP3-knockdown
   and control H1299 cells measured by z-score. The color means the z
   score of cluster analysis (C) KEGG enrichment analysis of
   differentially expressed genes (fold change > 1.5, p < 0.05) identified
   by RNA-seq in IGF2BP3-knockdown H1299 cells. The color means −log10 (p
   value) and the dot size means gene count of different pathways. D Gene
   Set Enrichment Analysis (GSEA) showing decreased fatty acid metabolism
   in IGF2BP3-knockdown cells. E, F Principal component analysis (PCA) and
   orthogonal partial least squares-discriminant analysis (OPLS-DA) of
   lipidomics data in IGF2BP3-knockdown H1299 cells. G Volcano plot of
   differential metabolites (fold change > 1.5, p < 0.05) in
   IGF2BP3-knockdown H1299 cells. H KEGG enrichment analysis of
   differential metabolites (fold change > 1.5, p < 0.05) identified by
   lipidomics in IGF2BP3-knockdown H1299 cells. The color means p value
   and dot size means gene count of pathways. I Heatmap of differential
   metabolites (fold change > 1.5, p < 0.05) between IGF2BP3-knockdown and
   control H1299 cells measured by z-score. The color means the z score of
   cluster analysis. J, K Neutral lipid levels detected by BODIPY 493/503
   staining in IGF2BP3-knockout brain tumors examined by microscope.
   Representative fluorescent images (J) and quantification (K) are shown.
   Data are presented as mean ± SD; n = 3. ***P < 0.001 (two-tailed
   t-test).

   To further examine the changes in metabolites related to lipid
   metabolism, we next performed lipomics analysis via LC-MS and carried
   out multivariate statistical analysis by principal component analysis
   (PCA) and orthogonal PLS-DA (OPLS-DA) (Fig. [123]3E, F). Lipomics
   analysis indicated that 15 upregulated and 133 downregulated
   metabolites in shIGF2BP3 H1299 cell lines (Fig. [124]3G). KEGG
   enrichment analysis indicated dysregulation in fatty acid biosynthesis
   and unsaturated fatty acid biosynthesis pathways (Fig. [125]3H).
   Heatmaps highlighted key metabolite changes, with IGF2BP3 knockdown
   reducing fatty acid, triglyceride, and diglyceride levels (Fig.
   [126]3I). Immunofluorescence staining of xenograft tumor in the brain
   confirmed decreased neutral lipid accumulation in IGF2BP3-knockout
   tumors (Fig. [127]3J, K).

   Functional assays further supported IGF2BP3’s role in lipid metabolism.
   IGF2BP3 knockdown reduced total triglycerides and cholesterol levels
   (Figs. [128]4A, B and S[129]4A-B). Then, we used the BODIPY 493/503
   probe to detect the neutral lipid levels, via fluorescent microscope.
   The result confirmed that IGF2BP3 knockdown significantly inhibits the
   neutral lipid levels (Figs. [130]4E, F and S[131]4E, F). Furthermore,
   we found that IGF2BP3 knockdown caused a decrease of lipid droplets
   (Figs. [132]4G, H and S[133]4G-H), using Nile red which is a dye
   commonly used to measure lipid droplets. Flow cytometry further
   confirmed reduced Nile red staining in IGF2BP3-knockdown cells (Figs.
   [134]4M, N and S[135]4M, N). Conversely, IGF2BP3 overexpression
   increased lipid accumulation, including triglycerides (Figs. [136]4C
   and S[137]4C), cholesterol (Figs. [138]4D and S[139]4D), neutral lipids
   (Figs. [140]4I, J and S[141]4I, J), and lipid droplets (Figs. [142]4K,
   L, O–P, S[143]4K, L and S[144]4O, P). These findings establish IGF2BP3
   as a critical regulator of lipid metabolism in NSCLC cells.

Fig. 4. IGF2BP3 regulates lipid metabolism in NSCLC cells.

   [145]Fig. 4
   [146]Open in a new tab

   A Total triglycerides assay kit showing total triglyceride levels were
   decreased in IGF2BP3-knockdown H1299 cells compared with control cells.
   Data are presented as mean ± SD; n = 3. B Total cholesterol assay kit
   showing total cholesterol levels were decreased in IGF2BP3-knockdown
   H1299 cells compared with control cells. Data are presented as
   mean ± SD; n = 3. C Total triglycerides assay kit showing total
   triglyceride levels were elevated in IGF2BP3-overexpressing H1299 cells
   compared with control cells. Data are presented as mean ± SD; n = 3. D
   Total cholesterol assay kit showing total cholesterol levels were
   elevated in IGF2BP3-overexpressing H1299 cells compared with control
   cells. Data are presented as mean ± SD; n = 3. E, F Neutral lipid
   levels detected by BODIPY 493/503 staining in IGF2BP3-knockdown H1299
   cells. Representative fluorescent images (E) and quantification (F) are
   shown. Data are presented as mean ± SD; n = 3. G, H Neutral lipid
   levels measured by Nile Red staining in IGF2BP3-knockdown H1299 cells.
   Representative fluorescent images (G) and quantification (H) are shown.
   Data are presented as mean ± SD; n = 3. I, J Neutral lipid levels
   detected by BODIPY 493/503 staining in IGF2BP3-overexpressing H1299
   cells. Representative fluorescent images (I) and quantification (J) are
   shown. Data are presented as mean ± SD; n = 3. K, L Neutral lipid
   levels measured by Nile Red staining in IGF2BP3-overexpressing H1299
   cells. Representative fluorescent images (K) and quantification (L) are
   shown. Data are presented as mean ± SD; n = 3. M, N Neutral lipid
   levels analyzed by Nile Red staining using flow cytometry in
   IGF2BP3-knockdown H1299 cells. Representative images (M) and
   quantification (N) are shown. Data are presented as mean ± SD; n = 3.
   O, P Neutral lipid levels analyzed by Nile Red staining using flow
   cytometry in IGF2BP3-overexpressing H1299 cells. Representative images
   (O) and quantification (P) are shown. Data are presented as mean ± SD;
   n = 3. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001 (two-tailed
   t-test).

IGF2BP3 stabilizes FASN mRNA to regulate lipid metabolism in NSCLC cells

   Given the role of IGF2BP3 in lipid production regulation, we sought to
   identify its potential binding targets in lipid metabolism. We carried
   out cDNA microarrays analysis, which demonstrated the expression of
   fatty acid metabolism-related genes in H1299 cells after IGF2BP3
   knockdown. We found that the key genes of fatty acid metabolism
   including REEP6, ALDOA, FASN, CRAT and CEL were markedly down-regulated
   after IGF2BP3 knockdown(Fig. [147]5A). Analysis of four independent
   IGF2BP3-RIP datasets from the GEO database identified 16 potential
   IGF2BP3-binding mRNAs (Fig. [148]5B). Overlapping these genes with
   fatty acid metabolism-related signatures highlighted FASN, a key enzyme
   in the endogenous lipogenesis pathway [[149]36], as a potential target
   of IGF2BP3-mediated RNA binding (Fig. [150]5C).

Fig. 5. IGF2BP3 stabilizes FASN mRNA to regulate lipid metabolism in NSCLC
cells.

   [151]Fig. 5
   [152]Open in a new tab

   A RT-qPCR analysis of fatty acid metabolism-related genes with cDNA
   microarrays in IGF2BP3-knockdown H1299 cells (n = 3). The color means
   the relative mRNA expression level of indicated genes. B Overlapping
   analysis of IGF2BP3-regulated genes detected by RIP-seq in
   [153]GSE90639, [154]GSE79147, [155]GSE230646, and [156]GSE231938 (fold
   change > 1.5, p < 0.05). C Overlapping analysis of genes from (A) and
   (B). D RIP-seq data showing potential binding of FASN mRNA to IGF2BP3
   according to [157]GSE231938 dataset. E RIP assay in
   IGF2BP3-overexpressing H1299 cells were verified by Western blotting. F
   RIP-qPCR analysis of FASN mRNA enrichment in IGF2BP3-overexpressing
   H1299 cells. Data are presented as mean ± SD; n = 3. G Pearson
   correlation analysis of IGF2BP3 and FASN expression in metastasis tumor
   tissue of LUAD patients according to [158]GSE198291 dataset (n = 312).
   H, I FASN mRNA expression evaluated by RT-qPCR in IGF2BP3-knockdown
   H1299 and H157 cells. Data are presented as mean ± SD; n = 3. J, K FASN
   protein expression evaluated by Western blotting in IGF2BP3-knockdown
   H1299 and H157 cells. L FASN mRNA half-life assessed by RT-qPCR in
   IGF2BP3-knockdown H1299 cells treated with 20 μg/mL actinomycin D for
   0, 1, 3, 6, 9, 16, and 24 hours. Data are presented as mean ± SD;
   n = 3. M FASN mRNA expression evaluated by RT-qPCR in shCtrl and
   shMETTL3 H1299 cells. Data are presented as mean ± SD; n = 3. N FASN
   and IGF2BP3 protein levels evaluated by Western blotting in shCtrl and
   shMETTL3 H1299 cells. O FASN mRNA half-life assessed by RT-qPCR in
   METTL3-knockdown H1299 cells treated with 20 μg/mL actinomycin D for 0,
   1, 3, 6, 9, 16, and 24 h. Data are presented as mean ± SD; n = 3.
   *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001 (two-tailed
   t-test).

   To validate this interaction, we analyzed the IGF2BP3-RIP seq data
   (Fig. [159]5D) and performed RIP-qPCR and confirmed that IGF2BP3
   directly binds to FASN mRNA in NSCLC cells (Fig. [160]5E, F). Analysis
   of the [161]GSE198291 dataset further demonstrated a significant
   positive correlation between IGF2BP3 and FASN mRNA expression in
   metastasis tumor tissue of LUAD patients (Fig. [162]5G). RT-qPCR and
   immunoblotting analyses revealed that FASN expression was significantly
   downregulated upon IGF2BP3 knockdown in NSCLC cells (Fig. [163]5H–K).
   Additionally, the half-life of FASN mRNA was markedly reduced in
   IGF2BP3-knockdown H1299 cells, indicating that IGF2BP3 stabilizes FASN
   mRNA (Fig. [164]5L). Previous studies have shown that IGF2BP3 is an m6A
   reading protein [[165]29, [166]37], we investigated whether IGF2BP3
   regulates FASN translation in an m6A-dependent manner. We established
   m6A-deficient NSCLC cells by knocking down METTL3 [[167]38], a crucial
   m6A methyltransferase, to measure the mRNA and protein levels of FASN.
   Our results revealed the down-regulation of FASN expression while
   IGF2BP3 remained unchanged (Fig. [168]5M, N). Consistent with this, the
   half-life of FASN mRNA was significantly reduced in METTL3-knockdown
   cells compared to controls (Fig. [169]5O). Furthermore, we detected the
   mRNA levels of FASN in stable IGF2BP3-konckdown and
   METTL3-overexpression NSCLC cells. Our results revealed that forced
   expression of METTL3 rescued the expression of FASN in
   IGF2BP3-deficient NSCLC cells (Fig. S[170]5A). These findings suggest
   that FASN is a direct target of IGF2BP3 in regulating lipid metabolism
   in NSCLC cells.

   As a downstream target of IGF2BP3, FASN may play a critical role in
   lipogenesis and metastasis. To assess this, we generated stable
   FASN-knockdown NSCLC cell lines and evaluated their functional
   phenotypes (Fig. S[171]5B-E). Our results showed that FASN knockdown
   reduced cell proliferation (Fig. S[172]5F-I), colony formation (Fig.
   S[173]5J–L), cell invasion (Figs. S[174]6A, B and S[175]7A, B) and
   migration (Fig. S[176]6C, D and S[177]7C, D). The flow cytometry
   analysis demonstrated that FASN knockdown increased cell apoptosis
   (Fig. S[178]5M–P). Subsequently, we also explored the role of FASN in
   lipogenesis in NSCLC cells. FASN knockdown decreased total
   triglycerides and cholesterol levels (Figs. S[179]6E, F and S[180]7E,
   F), reduced lipid droplet formation (Figs. S[181]6G, H, S[182]6K, L,
   S[183]7G, H, S[184]7K, L), and lowered neutral lipid levels (Figs.
   S[185]6I, J and S[186]7I, J). These results demonstrate that FASN is
   essential for lipid metabolism and invasiveness in NSCLC cells.

The IGF2BP3-FASN axis mediates lipid metabolism to promote brain colonization
in NSCLC

   To further elucidate the role of the IGF2BP3-FASN axis in NSCLC
   invasiveness and lipogenesis, we generated IGF2BP3-knockdown and
   FASN-overexpressing (shIGF2BP3+OE-FASN) H1299 and H157 cell lines. The
   mRNA level and protein expression were verified by RT-qPCR and
   immunoblotting (Fig. [187]6A–D). FASN overexpression significantly
   rescued the inhibition of cell proliferation (Fig. S[188]8A–D), colony
   formation (Fig. S[189]8E–G), migration (Fig. [190]6E–H), and invasion
   (Fig. [191]6I–L) induced by IGF2BP3 knockdown, while reducing apoptosis
   (Fig. S[192]8H–K). Consistently, we found that FASN overexpression
   effectively rescued IGF2BP3-knockdown mediated decreases in total
   triglycerides (Fig. [193]6M, O), total cholesterol (Fig. [194]6N, P),
   lipid droplets (Figs. [195]6Q–T, W, X and S[196]8N, O) and neutral
   lipid level (Figs. [197]6U, V and S[198]8L, M) in NSCLC cells. These
   findings confirm that FASN partly contributes to IGF2BP3-mediated lipid
   metabolism and invasiveness in NSCLC cells.

Fig. 6. IGF2BP3-FASN axis regulates invasiveness and lipid metabolism in
NSCLC cells.

   [199]Fig. 6
   [200]Open in a new tab

   A, B IGF2BP3 and FASN mRNA expression evaluated by RT-qPCR in
   IGF2BP3-knockdown and FASN-overexpressing H1299 and H157 cells. Data
   are presented as mean ± SD; n = 3. C, D IGF2BP3 and FASN protein levels
   evaluated by Western blotting in IGF2BP3-knockdown and
   FASN-overexpressing H1299 and H157 cells. E Wound healing assay showing
   cell migration ability in H1299 cells treated with IGF2BP3 knockdown
   and FASN overexpression for 24 h. Scale bars = 50 μm. F Quantification
   of cell migration in IGF2BP3-knockdown and FASN-overexpressing H1299
   cells from (E) using ImageJ. Data are presented as mean ± SD; n = 3. G
   Wound healing assay showing cell migration ability in H157 cells
   treated with IGF2BP3 knockdown and FASN overexpression for 24 h. Scale
   bars = 50 μm. H Quantification of cell migration in IGF2BP3-knockdown
   and FASN-overexpressing H157 cells from (G) using ImageJ. Data are
   presented as mean ± SD; n = 3. I Transwell invasion assay showing
   invasion ability in H1299 cells treated with IGF2BP3 knockdown and FASN
   overexpression for 24 h. Scale bars = 10 μm. J Quantification of cell
   invasion in IGF2BP3-knockdown H1299 cells from (I) using ImageJ. Data
   are presented as mean ± SD; n = 3. K Transwell invasion assay showing
   invasion ability in H157 cells treated with IGF2BP3 knockdown and FASN
   overexpression for 24 h. Scale bars = 10 μm. L Quantification of cell
   invasion in IGF2BP3-knockdown H157 cells from (K) using ImageJ. Data
   are presented as mean ± SD; n = 3. M Total triglycerides assay kit
   showing total triglyceride levels were rescued in IGF2BP3-knockdown and
   FASN-overexpressing H1299 cells compared with IGF2BP3-knockdown cells.
   Data are presented as mean ± SD; n = 3. N Total cholesterol assay kit
   showing total cholesterol levels were rescued in IGF2BP3-knockdown and
   FASN-overexpressing H1299 cells compared with IGF2BP3-knockdown cells.
   Data are presented as mean ± SD; n = 3. O Total triglycerides assay kit
   showing total triglyceride levels were rescued in IGF2BP3-knockdown and
   FASN-overexpressing H157 cells compared with IGF2BP3-knockdown cells.
   Data are presented as mean ± SD; n = 3. P Total cholesterol assay kit
   showing total cholesterol levels were rescued in IGF2BP3-knockdown and
   FASN-overexpressing H157 cells compared with IGF2BP3-knockdown cells.
   Data are presented as mean ± SD; n = 3. Q–T Neutral lipid levels
   analyzed by flow cytometry in IGF2BP3-knockdown and FASN-overexpressing
   H1299 and H157 cells stained with Nile Red. Representative images (Q,
   S) and quantification (R, T) are shown. Data are presented as
   mean ± SD; n = 3. U, V Neutral lipid levels detected by BODIPY 493/503
   staining in IGF2BP3-knockdown and FASN-overexpressing H1299 cells.
   Representative fluorescent images (U) and quantification (V) are shown.
   Data are presented as mean ± SD; n = 3. W, X Neutral lipid levels
   measured by Nile Red staining in IGF2BP3-knockdown and
   FASN-overexpressing H1299 cells. Representative fluorescent images (W)
   and quantification (X) are shown. Data are presented as mean ± SD;
   n = 3. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001 (two-tailed
   t-test).

   Clinically, FASN was upregulated in NSCLC-derived leptomeningeal and
   brain metastatic tumors compared to primary tumors in the [201]GSE83132
   dataset (Fig. [202]7A, B). Kaplan-Meier survival analysis revealed that
   high FASN expression correlates with poor prognosis in metastatic lung
   cancer patients (Fig. [203]7C). To validate these findings in vivo, we
   established a mouse stereotactic brain injection model using H1299-luc
   cells with IGF2BP3 knockout (sgIGF2BP3-luc) or FASN rescue
   (sgIGF2BP3+OE-FASN-luc), and verified the protein level by
   immunoblotting (Fig. [204]7D). H1299-luc cells were injected into the
   cortex of 4-week-old nude mice and the body weight was measured every
   three days (Fig. S[205]8P). The tumors were imaged in vivo by IVIS
   system every nine days and plotted a growth curve. Next, the mice
   brains were harvested and imaged on day 27. Consistent with the in
   vitro study, FASN overexpression partly reversed the reduced tumor
   burden in the brain caused by IGF2BP3 knockout (Fig. [206]7E–H). IHC
   staining confirmed reduced IGF2BP3 and increased FASN expression in
   brain colonization tumor, with FASN rescue restoring MMP9 expression
   (Fig. [207]7I–J). Immunofluorescence staining further demonstrated that
   FASN overexpression rescued neutral lipid production in brain
   colonization tumor (Fig. [208]7K, L). Collectively, these results
   highlight that the IGF2BP3-FASN axis is a potential therapeutic target
   for regulating the growth of NSCLC cells in the brain through lipid
   metabolism.

Fig. 7. The IGF2BP3-FASN axis mediates lipid metabolism to promote brain
colonization in NSCLC.

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

   A FASN expression was elevated in LLC1-derived leptomeningeal
   metastatic cells (Lep) compared with primary cells according to
   [211]GSE83132 dataset. Data are presented as mean ± SD; n = 3. B FASN
   expression was elevated in PC9-derived brain metastatic cells (Brm)
   compared with primary cells according to [212]GSE83132 dataset. Data
   are presented as mean ± SD; n = 3. C Overall survival of metastatic
   lung cancer patients was calculated with the log-rank test and
   stratified by high or low FASN expression (high: top 50%, n = 27; low:
   bottom 50%, n = 26; [213]GSE30219). D IGF2BP3 and FASN protein
   expression detected by Western blotting in IGF2BP3-knockout and
   FASN-overexpressing H1299-luc cells. E, F H1299-luc cells (5 × 10^5)
   with IGF2BP3 knockout and FASN overexpression were injected into
   BALB/c-nu/nu mice via stereotactic brain injection (n = 6 per group).
   Representative bioluminescent images of NSCLC xenograft tumor in the
   brain on day 27 are shown in (E). Quantification of bioluminescence on
   day 27 is shown in (F). Data are presented as mean ± SD; n = 6. G, H
   Representative bioluminescent images of brain tumors on day 27 were
   captured by the Neurostar IVIS preclinical in vivo imaging system (G).
   Quantification of bioluminescence in brain tumors on day 27 is shown in
   (H). Data are presented as mean ± SD; n = 6. I H&E staining and IHC
   analysis of IGF2BP3 and MMP9 in representative brain tumors from
   IGF2BP3-knockout and FASN-overexpressing mice. Scale bar = 20 μm. J
   Quantification of IGF2BP3 and MMP9 expression levels from IHC analysis
   in (I) using Image Pro Plus. Data are presented as mean ± SD; n = 3. K,
   L Neutral lipid levels detected by BODIPY 493/503 staining in
   representative brain tumors from IGF2BP3-knockout and
   FASN-overexpressing mice. Representative fluorescent images (K) and
   quantification (L) are shown. Data are presented as mean ± SD; n = 3.
   *P < 0.05; **P < 0.01; ****P < 0.0001 (two-tailed t-test).

Discussion

   In this study, we elucidated the critical role of lipid metabolism in
   driving the growth of NSCLC cells in the brain through the IGF2BP3-FASN
   axis. IGF2BP3, initially identified as an oncofetal protein
   [[214]39–[215]41], is overexpressed in various cancers and has been
   implicated in tumor proliferation, drug resistance, and metastasis
   [[216]29, [217]42–[218]44]. NSCLC patients frequently develop brain
   metastases during disease progression, yet current therapeutic options
   remain limited, underscoring the need for novel therapeutic targets.
   Here, we demonstrated that IGF2BP3 is a potential therapeutic target
   for NSCLC brain colonization, acting through FASN-mediated lipid
   metabolism. IGF2BP3 is upregulated in NSCLC brain metastases and
   positively correlates with key epithelial-mesenchymal transition (EMT)
   genes, as supported by bioassays and bioinformatics analyses. Elevated
   IGF2BP3 expression is associated with poor clinical prognosis and
   advanced metastatic stages, highlighting its important role in tumor
   metastasis and its potential as a therapeutic target. Consistent with
   our findings, recent studies have linked IGF2BP3 to brain metastasis in
   breast cancer [[219]34], further supporting its role in tumor
   dissemination to the brain. However, the mechanisms that IGF2BP3
   mediates brain colonization of NSCLC cells remain incompletely
   understood.

   Our functional studies revealed that targeting IGF2BP3 significantly
   inhibits invasiveness in vitro and reduces NSCLC tumor burden in the
   brain in vivo. IGF2BP3 knockdown suppressed NSCLC cell proliferation,
   migration, and invasion while promoting apoptosis. Using a mouse
   stereotactic brain injection model, we found that IGF2BP3 knockout
   significantly reduced NSCLC tumor burden in the brain, whereas IGF2BP3
   overexpression exacerbated brain colonization. These findings suggest
   that IGF2BP3 is a promising therapeutic target for NSCLC brain
   metastasis.

   To uncover the mechanisms by which IGF2BP3 promotes the growth of NSCLC
   cells in the brain, we employed RNA sequencing and lipidomics analyses.
   Gene enrichment analysis highlighted metabolic and lipogenesis pathways
   as key downstream targets of IGF2BP3. The brain microenvironment
   imposes unique metabolic constraints on cancer cells, and recent
   studies have demonstrated that metabolic adaptations, such as de novo
   serine synthesis and acetate utilization, are critical for tumor
   survival in the brain. Lipid metabolism, in particular, plays a pivotal
   role in brain metastasis. For example, Ngo and colleagues found that
   tumor cells require de novo serine synthesis to proliferate in the
   brain microenvironment in breast cancer brain metastases [[220]9],
   which has reduced amino acid levels relative to plasma [[221]45,
   [222]46]. Mashimo and colleagues demonstrated that brain metastasis
   showed special acetate adaptation for meeting the high biosynthetic and
   bioenergetic demands in glioblastoma [[223]47]. It was reported that
   fatty acid synthesis is required for breast cancer brain metastasis
   [[224]17]. Consistently, lipid synthesis was more active in brain
   metastasis in breast cancer [[225]48]. These studies suggested that the
   brain microenvironment can impose metabolic constraints on cancer cells
   and that identifying metabolic limitations could inform strategies to
   overcome brain metastasis, especially lipid metabolism. Our lipidomics
   analysis revealed that IGF2BP3 knockdown downregulates various lipid
   metabolites, including fatty acids and triglycerides. Recent study
   highlighted that IGF2BP3 regulate Stearoyl-CoA Desaturase (SCD), a key
   enzyme in lipogenesis, to affect the metabolic pathway in cervical
   cancer brain metastasis [[226]32]. Consistent with these findings,
   IGF2BP3 depletion reduced neutral lipid accumulation, total
   triglycerides, cholesterol levels, and lipid droplet formation in NSCLC
   cells. These results underscore the role of IGF2BP3 in regulating
   lipogenesis and fatty acid metabolism in NSCLC.

   We further investigated the mechanism by which IGF2BP3 modulates fatty
   acid metabolism. Overlapping RNA-seq and RIP-seq data identified FASN,
   a key enzyme in lipogenesis, as a downstream target of IGF2BP3. FASN is
   central to fatty acid metabolism [[227]49] and mediates lipogenesis to
   facilitated the growth of a variety of human cancers [[228]13,
   [229]50]. Recent study also showed that FASN activating fatty acid
   metabolism to promote cancer metastasis in male breast cancer
   [[230]16]. In our study, FASN was essential for IGF2BP3-mediated lipid
   metabolism and brain colonization. IGF2BP3 knockdown significantly
   reduced FASN expression at both mRNA and protein levels and decreased
   FASN mRNA stability. Mechanistically, IGF2BP3 directly binds to FASN
   mRNA, stabilizing it in an RNA-binding-dependent manner. Clinically,
   FASN is upregulated in leptomeningeal and brain metastatic tumors and
   correlates with poor prognosis in NSCLC patients. These findings align
   with previous studies linking FASN to metastasis in osteosarcoma and
   colorectal cancer [[231]51, [232]52].

   To determine whether FASN is the crucial downstream effector of
   IGF2BP3-mediated brain metastasis, we overexpressed FASN in
   IGF2BP3-knockdown NSCLC cells. FASN overexpression partly rescued the
   phenotypes of reduced growth, invasiveness, and neutral lipid
   accumulation caused by IGF2BP3 depletion. Furthermore, we generated a
   mouse stereotactic brain injection model based on NSCLC cells. FASN
   overexpression effectively alleviated IGF2BP3-dependent brain
   colonization and restored lipid droplet formation in a mouse
   stereotactic brain injection model. These results suggest that the
   IGF2BP3-FASN axis plays a critical role in the growth of NSCLC cells in
   the brain microenvironment. While FASN overexpression partially rescued
   IGF2BP3-knockdown effects on lipid metabolism and tumor burden in
   brain, the incomplete restoration implies additional downstream
   mediators or mechanisms may also contribute. Furthermore, it is
   undeniable that our model focuses specifically on the later stages of
   metastasis (colonization of NSCLC cells in the brain microenvironment),
   rather than the entire metastatic process. In the future, further
   investigation and validation of these findings using spontaneous
   metastasis models (intracardiac or orthotopic injection) will enrich
   and deepen the understanding of the role of the IGF2BP3-FASN axis in
   brain metastasis of NSCLC. Notwithstanding these limitations, this
   study does suggest the potential of IGF2BP3 as a drive gene for NSCLC
   cells to grow in the brain.

   In summary, our study identifies IGF2BP3 as a key regulator of the
   growth of NSCLC cells in the brain microenvironment through lipid
   metabolism. IGF2BP3 is upregulated in NSCLC brain metastases and
   promotes FASN-mediated neutral lipid accumulation. Mechanistically,
   IGF2BP3 binds to FASN mRNA, enhancing its stability in an
   RNA-binding-dependent manner. The IGF2BP3-FASN axis emerges as a
   critical regulator of brain colonization, linking lipid metabolism to
   tumor dissemination. These findings highlight IGF2BP3 as a potential
   therapeutic target for inhibiting the growth of NSCLC cell in the
   brain.

Materials and methods

Cell culture

   The H1299, H157 and HEK-293T cell lines were obtained from the American
   Type Culture Collection (ATCC, USA). The H1299 and H157 cell lines were
   cultured in RPMI-1640 (Gibco, USA) supplemented with 10% FBS (Gibco,
   USA) and 1% penicillin-streptomycin (PS, Gibco, USA). HEK-293T cells
   were cultured in DMEM (Gibco, USA) with 10% FBS and 1% PS. All cell
   lines were cultivated at 37 °C under 5% CO₂. To minimize phenotypic
   changes, cells were aliquoted into multiple vials upon receipt and
   routinely tested for mycoplasma contamination using the Mycoplasma
   Stain Assay Kit (C0296, Beyotime, China). Cell line authentication was
   performed using STR analysis with the GenePrint 10 System (B9510,
   Promega, USA) following the manufacturer’s instructions.

Plasmid construction, cell transfection, and virus infection

   Stable knockout of target genes was achieved using the lentiCRISPRv2
   system, while stable knockdown was accomplished via lentiviral delivery
   of specific short-hairpin RNAs (shRNAs). Primer sequences for vector
   construction are listed in Table S[233]1.

   For overexpression studies, human IGF2BP3 cDNA ([234]NM_006547) and
   FASN cDNA ([235]NM_004104.5) were cloned into the pCDNA3.1 lentiviral
   vector (CD510B-1, System Biosciences).

Antibodies

   The following antibodies were used: anti-IGF2BP3 (WB 1:1000; ab179807,
   Abcam, USA), anti-IGF2BP3 (IHC 1:100; EM1701-15, HuaBio, China),
   anti-FASN (WB 1:500; IHC 1:100; 66591-1-Ig, Proteintech, China),
   anti-β-actin (WB 1:1000, #4967, Cell Signaling Technology, USA),
   anti-MMP9 (IHC 1:500; GB11132, Servicebio, China), and anti-Cytokeratin
   18 (IF 1:200; GB11232, Servicebio, China).

Western blotting

   Cells were lysed in 100 μL modified RIPA buffer (P0013B, Beyotime,
   China) containing protease inhibitors (#11836153001, Roche,
   Switzerland) and phosphatase inhibitors (Bimake, USA). Protein
   concentrations were determined using the BCA protein assay kit (23252,
   ThermoFisher, USA). Proteins were separated by 8–10% SDS-PAGE and
   transferred to polyvinylidene fluoride (PVDF) membranes (Millipore,
   USA). Membranes were blocked with 5% BSA and probed with primary and
   secondary antibodies. Target proteins were visualized using a
   chemiluminescence imaging system (1708280, Bio-Rad, USA).

Immunohistochemistry

   Tissue slides were deparaffinized, rehydrated, and subjected to antigen
   retrieval in sodium citrate buffer. Tumor sections were blocked with 5%
   goat serum containing 0.1% Triton X-100 and 3% H₂O₂ in PBS for 60 min
   at room temperature, followed by incubation with primary antibodies at
   4 °C overnight. IHC staining was performed using horseradish peroxidase
   (HRP)-conjugated secondary antibodies and DAB detection. Nuclei were
   counterstained with DAPI, and images were acquired using an imaging
   system (AMAFD1000, ThermoFisher, USA).

Cell proliferation assays

   For the CCK-8 assay, cells were seeded at 1000 cells/well in 96-well
   plates. Cell viability was assessed at 0, 24, 48, 72, 96, and 120 h
   using the CCK-8 reagent. Absorbance was measured at 450 nm using a
   microplate reader (FLUOstar, Omega, USA) after a 2-h incubation at
   37 °C. For cell counting assays, cells were seeded at 5 × 10^4
   cells/well in 6-well plates, and cell numbers were counted at the
   indicated time points.

Colony formation assay

   Cells were seeded in 6-well plates at densities of 500 cells/well
   (H1299) or 1 × 10^3 cells/well (H157) and cultured for 10 days.
   Colonies were fixed with 4% paraformaldehyde, stained with crystal
   violet (Beyotime, China), and imaged.

Cell apoptosis assay

   Apoptosis was assessed using the Annexin V-PI Apoptosis Detection Kit I
   (BB-4101, BestBio, China). Briefly, cells were collected, centrifuged
   at 1000 × g for 3 min, and resuspended in 400 μL annexin V-FITC binding
   buffer containing 5 μL annexin V-FITC and 10 μL propidium iodide.
   Samples were analyzed by flow cytometry (CytoFLEX, Beckman, USA) and
   quantified using FlowJo V10 software.

Wound healing assay

   Cells were seeded in 6-well plates and grown to 90% confluence. A
   sterile pipette tip was used to create a scratch, and detached cells
   were removed by washing with PBS. Cells were incubated in 2% FBS medium
   for 24 h, and images were captured at 0 and 24 h. The rate of wound
   healing = [(the wound width of 0 h–48 h)/0 h wound width] × 100%.

Transwell invasion

   Transwell chambers (Thermo Fisher Scientific) were used for invasion
   assays. Cells were seeded in the upper chamber in 100 μL of medium
   containing 10% FBS, while the lower chamber contained medium with 20%
   FBS as a chemoattractant. After 24 h, non-invading cells were removed,
   and invading cells were fixed with 4% paraformaldehyde, stained with 1%
   crystal violet, and imaged using an EVOS M5000 Fluorescence Microscope.

Measurement of TG and TC

   Total triglycerides (TG) and total cholesterol (TC) levels were
   measured using the Total Triglycerides Assay Kit (BB-47436, Bestbio)
   and Total Cholesterol Assay Kit (BB-47435, Bestbio), respectively.
   Cells were washed with isotonic buffer, lysed in PBS containing 3%
   Triton X-100, and centrifuged. TG or TC assay reagents were added to
   the lysates, and absorbance was measured at 520 nm using a microplate
   reader. Concentrations were calculated according to the manufacturer’s
   protocols.

Bodipy and Nile Red assays

   Cells were fixed with 4% formaldehyde. For BODIPY staining, cells were
   incubated in PBS for 30 min. For Nile Red staining, cells were
   incubated in PBS for 1 h. After washing with PBS, nuclei were
   counterstained with DAPI. Fluorescence was visualized using an
   FV-1000/ES confocal microscope (Olympus, Japan) or quantified by flow
   cytometry.

Quantitative RT-qPCR assay

   Total RNA was extracted using Trizol reagent (15596026, Invitrogen,
   USA) and treated with DNase I to ensure RNA quality. cDNA synthesis and
   quantitative PCR were performed using SYBR PCR Master Mix (Yeasen,
   China). Primer sequences are listed in Table S[236]2.

Metabolomics analysis

   H1299 and H1299 shIGF2BP3 cells (1 × 10^7 cells) were homogenized in
   1 mL of a methanol:ddH[2]O solution (4:1, v/v) and centrifuged at
   12,000×g at 4 °C for 10 min. The supernatant was collected, dried using
   a vacuum concentrator, and reconstituted in 300 μL of acetonitrile:0.1%
   formic acid (FA) solution (1:9, v/v, pre-chilled at −20 °C). After
   filtration through a 0.22 μm membrane, the filtrate was transferred to
   a detection vial. Chromatographic separation was performed using a
   Thermo Ultimate 3000 system equipped with an ACQUITY UPLC® HSS T3
   column (150 × 2.1 mm, 1.8 μm, Waters) maintained at 40 °C. The
   autosampler temperature was set to 8 °C. Gradient elution was achieved
   using 0.1% FA in water (solvent C) and 0.1% FA in acetonitrile (solvent
   D) or 5 mM ammonium formate in water (solvent A) and acetonitrile
   (solvent B) at a flow rate of 0.25 mL/min. Each sample (2 μL) was
   injected after equilibration. The gradient program was as follows:
   0–1 min, 2% B/D; 1–9 min, 2–50% B/D; 9–12 min, 50–98% B/D; 12–13.5 min,
   98% B/D; 13.5–14 min, 98–2% B/D; 14–20 min, 2% D (positive mode) or
   14–17 min, 2% B (negative mode). ESI-MSn experiments were conducted on
   a Thermo Q Exactive mass spectrometer with spray voltages of 3.8 kV
   (positive mode) and −2.5 kV (negative mode). Sheath and auxiliary gases
   were set to 45 and 15 arbitrary units, respectively, and the capillary
   temperature was maintained at 325 °C. The Orbitrap analyzer scanned a
   mass range of m/z 81–1000 at a resolution of 70,000. Data-dependent
   acquisition (DDA) MS/MS experiments were performed using higher-energy
   collisional dissociation (HCD) with a normalized collision energy of
   30 eV. Dynamic exclusion was applied to remove redundant data. LC-MS
   detection and analysis were performed by Wininnovate Bio, LTD
   (Shenzhen, China).

RNA immunoprecipitation sequencing

   RNA immunoprecipitation (RIP) assays were performed using the Magna
   RIP™ RNA-Binding Protein Immunoprecipitation Kit (17–700, Merck
   Millipore, USA) according to the manufacturer’s protocol. Briefly,
   stable IGF2BP3-overexpressing H1299 cells were lysed in complete RIP
   lysis buffer. Anti-IGF2BP3 mouse monoclonal antibody (EM1701-15,
   HuaBio, China) and anti-IgG mouse monoclonal antibody were conjugated
   to magnetic beads to immunoprecipitate RNA-protein complexes. Enriched
   mRNAs were validated by RT-qPCR using RIP-specific primers.

RNA stability assay (half-life detection)

   RNA stability was assessed by treating cells with 20 μg/mL actinomycin
   D (Act-D, HY-17559, MCE, USA). Cells were collected at 0, 1, 3, 6, 9,
   16, and 24-h post-treatment, and RNA was isolated for RT-qPCR analysis.

Animal experiments

   All animal experiments were approved by the Animal Ethics Committee of
   Sun Yat-Sen University and conducted in accordance with the Guide for
   the Care and Use of Laboratory Animals. Four-week-old male BALB/c nude
   mice (18 g, SPF grade) were obtained from the Experimental Animal
   Center of Sun Yat-sen University (Guangdong, China).

   For the subcutaneous tumor model, H1299 cells (5 × 10^6) were injected
   into the nude mice. Tumor growth was monitored for 27 days until
   harvested. At the endpoint, mice were euthanized, and subcutaneous
   tumor were harvested for histological analysis.

   For the stable sgIGF2BP3 stereotactic brain tumor model, sgCtrl or
   sgIGF2BP3 H1299 cells (5×10⁵) were injected into the cortex of nude
   mice. For the IGF2BP3 overexpression model, OE-Ctrl or OE-IGF2BP3 H1299
   cells (5 × 10^5) were injected. For rescue experiments, sgCtrl,
   sgIGF2BP3, or sgIGF2BP3&OE-FASN H1299 cells (5 × 10^5) were injected.
   Tumor growth was monitored using the Neurostar IVIS preclinical in vivo
   imaging system every 9 days lasted for a total of 27 days. At the
   endpoint, mice were euthanized, and brains were harvested for
   histological analysis.

Bioinformatics analysis

   Gene expression profiles (TCGA-LUAD, GSE11049, [237]GSE83132,
   [238]GSE33532, [239]GSE198291, [240]GSE90639, [241]GSE79147,
   [242]GSE230646, [243]GSE231938, and [244]GSE30219) were downloaded from
   the Gene Expression Omnibus (GEO) database and The Cancer Genome Atlas
   (TCGA, [245]http://cancergenome.nih.gov). The online database R2
   Genomics Analysis and Visualization Platform
   ([246]https://hgserver1.amc.nl/cgi-bin/r2/main.cgi) was applied to
   determine the correlation of the related genes in [247]GSE33532.
   Pearson analysis was applied to analyze the correlation of the related
   gene in [248]GSE198291. Differentially expressed genes (fold
   change > 1.5, p < 0.05) and RIP data (fold change > 2, p < 0.05) were
   analyzed. Kaplan-Meier Plotter ([249]https://kmplot.com/analysis/) was
   used to determine the clinical survival of the related genes (cutoff:
   50%; p < 0.05). KEGG pathway enrichment analysis was performed using
   DAVID Bioinformatics Resources 6.8 ([250]http://david.ncifcrf.gov).
   Metabolomics data were enriched and differential metabolites (Fold
   change > 1.5; p < 0.05) were analyzed using the MetaboAnalyst database
   ([251]https://www.metaboanalyst.ca/home.xhtml).

Statistical analysis

   Statistical analysis was performed by GraphPad Prism 8.0 software
   (GraphPad, Inc., La Jolla, CA, USA) and quantitative data were
   presented as the mean ± SD unless stated otherwise. Statistical
   differences were assessed using a two-tailed Student’s t-test. The
   sample sizes (n) and probability (p) values for each experiment were
   indicated in detail in figure legends. Survival curves were constructed
   using the Kaplan–Meier method and analyzed by the log-rank test. The
   correlation analysis was assessed by Pearson’s correlation analysis.
   The detail of colors and the node size for indicated graph were
   indicated in figure legends. To represent the results as a heat map,
   the z-score values were shown with the color intensity, and the
   remaining values were ranked according to their relative values. The
   graphic abstract was drawn by Figdraw ([252]https://www.figdraw.com).
   The schematics of mice cortex stereotactic injection and in vivo
   imaging system was drawn by Biorender ([253]https://biorender.com).
   Differences between values were considered statistically significant
   when *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Supplementary information

   [254]original western blot^ (683.9KB, pdf)
   [255]Supplementary materials^ (144.3MB, docx)

Acknowledgements