Protein lipidation in the tumor microenvironment: enzymology, signaling pathways, and therapeutics

Xu, Mengke; Xu, Bo

doi:10.1186/s12943-025-02309-7

Review
Open access
Published: 07 May 2025

Protein lipidation in the tumor microenvironment: enzymology, signaling pathways, and therapeutics

Mengke Xu¹ &
Bo Xu¹

Molecular Cancer volume 24, Article number: 138 (2025) Cite this article

769 Accesses
1 Altmetric
Metrics details

Abstract

Protein lipidation is a pivotal post-translational modification that increases protein hydrophobicity and influences their function, localization, and interaction network. Emerging evidence has shown significant roles of lipidation in the tumor microenvironment (TME). However, a comprehensive review of this topic is lacking. In this review, we present an integrated and in-depth literature review of protein lipidation in the context of the TME. Specifically, we focus on three major lipidation modifications: S-prenylation, S-palmitoylation, and N-myristoylation. We emphasize how these modifications affect oncogenic signaling pathways and the complex interplay between tumor cells and the surrounding stromal and immune cells. Furthermore, we explore the therapeutic potential of targeting lipidation mechanisms in cancer treatment and discuss prospects for developing novel anticancer strategies that disrupt lipidation-dependent signaling pathways. By bridging protein lipidation with the dynamics of the TME, our review provides novel insights into the complex relationship between them that drives tumor initiation and progression.

Introduction

The tumor microenvironment (TME) encompasses a complicated network of dynamic interactions among tumor cells, stromal cells, vasculature, immune cells, extracellular matrix, and secreted factors [1]. Tumor initiation and progression are influenced not only by the intrinsic properties of tumor cells, but also by the multifaceted interactions within the TME that continuously shape tumor behavior [2]. Recent studies have shown that post-translational modifications (PTMs), particularly protein lipidation, play crucial roles in TME functions [3, 4]. Lipidation is the covalent attachment of lipid groups to proteins that increases hydrophobicity, facilitating membrane interactions that influence protein localization, signal transduction, and cellular interactions [5]. It includes S-prenylation, N-myristoylation, S-palmitoylation, and other less common types, such as O-palmitoylation, N-palmitoylation, C-terminal cholesterol esterification, and GPI anchor attachment (Fig. 1). This review focuses on three major lipidation modifications: S-prenylation, S-palmitoylation, and N-myristoylation. N-myristoylation is an irreversible lipidation involving the attachment of the 14-carbon fatty acid myristate to the N-terminal glycine of a substrate, catalyzed by N-myristoyltransferase (NMT) (Fig. 2a) [6, 7]. This modification typically occurs cotranslationally. During the protein synthesis, the initiator methionine is often removed by the methionine aminopeptidase 2(MetAP2), exposing a glycine residue at the N-terminus. The NMT then transfers the myristoyl group from myristoyl-coenzyme A (CoA) to this glycine residue [8]. In contrast, S-palmitoylation is a reversible modification that involves the attachment of palmitate, a 16-carbon saturated fatty acid (palmitate), to cysteine residues via a thioester bond, catalyzed by zinc-finger and aspartate–histidine–histidine–cysteine (ZDHHC) type (ZDHHC1 - 24, without ZDHHC10) family proteins. S-palmitoylation occurs in two steps: first, the ZDHHC enzymes autopalmitoylate at the DHHC cysteine residue, and then the palmitate is transferred to the cysteine residue of the target protein [9] (Fig. 2b). S-prenylation, also irreversible [10], refers to the covalent attachment of either a farnesyl (15-carbon) or, more commonly, a geranylgeranyl (20-carbon) group to the C-terminal cysteine residues, which must be part of a C-terminal cysteine-aliphatic-aliphatic-X (CAAX) box motif, of proteins, a reaction mediated by farnesyltransferase (FTase) and geranylgeranyltransferases I, II, and III (GGTase) [11, 12] (Fig. 2c). More importantly, many proteins undergo dual lipid modifications [13]; dual lipidation can occur in two scenarios: N-myristoylation with S-palmitoylation and S-prenylation with S-palmitoylation. The mechanism is two irreversible modifications occurring initially, followed by S-palmitoylation. The hydrophobicity conferred by the first modification is insufficient for precise localization, whereas the subsequent S-palmitoylation ensures accurate protein targeting.

Recent discoveries have shown that inhibitors targeting lipidation-related enzymes exhibit potent antitumor properties [14]. These inhibitors interfere with oncogenic signal transmission and cellular survival pathways, causing metabolic dysregulation [15, 16], cell cycle arrest [17], and programmed cell death [18,19,20]. They also increase the sensitivity of tumor cells to immune-driven attacks [21]. Several related drugs have been researched in the clinic [22,23,24].

Although some aspects of lipidation modifications in the TME have been revealed, no review has been published thus far to summarize the detailed interactions between them. Our review discusses recent progress in protein lipidation, focusing on N-myristoylation, S-palmitoylation, and S-prenylation. Our review synthesizes all relevant studies on lipidation in tumors and the TME, comprehensively elucidating the complex and critical regulatory interactions of lipidation within the TME. We provide an overview of current lipidation-related therapeutics in cancer treatment. Through this work, we aim to offer novel insights and recommendations for understanding disease mechanisms and advancing therapeutic strategies in oncology.

Molecular mechanisms of lipidation

Mechanisms of lipidation

N-myristoylation

N-myristoylation, catalyzed by NMT1/NMT2 [25], attaches a myristoyl group to Gly2 of nascent proteins. Despite shared catalytic folds (GNAT domains) and overlapping substrate specificity, NMT1 exhibits broader substrate range [26] and is overexpressed in cancers [27], linking it to tumorigenesis and embryonic development [28, 29] (Fig. 2a).

Substrate recognition depends on a conserved binding cleft: the N-terminal glycine (Gly2) and adjacent residues (e.g., Cys3) are anchored via hydrophobic pockets and electrostatic interactions (e.g., Asp183/185/471), while Myr-CoA binding triggers conformational closure to form a stable ternary complex. Competition with NatA in ribosomal exit tunnels dynamically regulates protein acetylation versus N-myristoylation, expanding functional diversity. Evolutionarily, the N-terminal pentapeptide (residues 2–6) encodes a dual-modification "code", where Cys3 - 5 positions enable sequential N-myristoylation-palmitoylation synergism [26]. Clinically, N-myristoylation stabilizes oncoproteins (e.g., Src kinases) and modulates immune responses, making NMT1 a therapeutic target in TME [8, 30, 31].

S-palmitoylation

ZDHHC family enzymes catalyze S-palmitoylation through their conserved DHHC-cysteine-rich structural domains and at least four transmembrane structural domains [32], whose crystal structures (e.g., human ZDHHC20 and zebrafish ZDHHC15) show that the transmembrane structural domains form a hydrophobic lumen that selects for the length of the acyl chain (C14:0 to C22:0, usually C16:0) through residues in the cavity [33, 34]. Catalysis follows a two-step ping-pong mechanism: first, the conserved Cys of the DHHC motif reacts with palmitoyl-CoA to generate an acyl-enzyme intermediate (auto-S-palmitoylation); subsequently, the acyl group is transferred to a near-membrane Cys residue of the substrate protein.The substrate selectivity of the ZDHHC enzyme is determined by the subcellular localization because it dictates where enzymes and substrates can interact, the specific structural domains (such as the anchor protein repeat motif of ZDHHC13/17 [35] or PDZ-binding motif of ZDHHC5 [36]) and substrate sequence features (e.g., N-terminal N-myristoylation sequence of near-membrane Cys) synergistically [37]. The Golgi membrane is its primary active site [38], This modification dynamically regulates protein-membrane interactions by enriching targets (e.g., Fas, death receptor 4 (DR4)) in lipid rafts [39, 40], maintaining Golgi-PM distribution [41], and mediating lysosomal sorting via TMEM55B-dynein complexes [42].

The reversibility of this modification is mediated by acyl protein thioesterases (APT1/2, PPT1) and α/β hydrolases (ABHD10, ABHD17 A-C) [37], which enable the protein to switch palmitoylated/de-palmitoylated states on a second to hourly scale [43], dynamically modulating membrane affinity, localization, stability, and signal transduction functions [44].The results demonstrate the spatiotemporal plasticity of lipid modification in cellular response (Fig. 2b).

S-prenylation

S-prenylation involves the attachment of a farnesyl or geranylgeranyl group to cysteine residues within a CAAX motif, which determines the type of S-prenylation [45]. FTase catalyzes farnesylation, while GGTase mediates geranylgeranylation. GGTase exists in three isoforms; FTase and GGTase I share an α subunit but differ in β subunits for distinct substrate specificities [46]. Differential electrostatic potentials on their β subunits dictate substrate preference: FTase's β subunit favors serine/methionine at the "X" position of CAAX motif, whereas GGTase-I prefers leucine, these enzymes do not manifest mutually exclusive substrate specificity [47], some proteins can be processed by GGTase-I when FTase is inhibited [45, 48, 49]. A reasonable explanation is FTase favors Farnesyl pyrophosphate (FPP, 15 carbons) because its shorter chain positions the diphosphate closer to the catalytic Zn²⁺, while GGTase-I prefers Geranylgeranyl pyrophosphate (GGPP 20 carbons) due to deeper insertion into its hydrophobic pocket. FPP binds weakly to GGTase-I (330-fold lower affinity than GGPP), but its diphosphate still partially aligns with the active site, allowing limited catalysis [50]. GGTase-II specifically modifies Rab proteins with double-cysteine motifs for dual geranylgeranylation, which is essential for vesicle transportation; the geranylgeranylation of Rab occurs through two distinct mechanisms. First, Rab escort proteins (REPs) initially bind to Rab, facilitating its subsequent interaction with GGTase-II. In the second mechanism, REPs first associate with GGTase-II, and this complex then binds to Rab [51,52,53,54,55,56,57]. GGTase III modifies specific proteins, such as FBXL2 and Ykt6 [58, 59].

Indeed, S-prenylation involves three significant steps before membrane localization: preliminary S-prenylation, proteolytic cleavage, and carboxymethylation [60,61,62] (Fig. 2c). Initially, S-prenylation transports proteins to the ER. Secondly, the AAX residues are cleaved from the CAAX motif of the protein by Ras converting enzyme 1 (Rce1), which makes the protein more receptive to membrane binding [63]. Finally, isoprenylcysteine carboxyl methyltransferase (ICMT) methylates prenylated cysteine, further neutralizing its negative charge and enhancing membrane affinity [64, 65]. The modified protein is then targeted to membranes and dynamically distributed among distinct membrane regions (PDE6δ) [66], guanine nucleotide dissociation inhibitors (GDIs) [67], and Smg GDS [68]. Other dissociation factors, such as Arl2 and Arl3 and GDI displacement factors [67], are said to release these proteins from solubilizing factors and relocate them to membranes. KRAS4B exemplifies a unique targeting mechanism, its PM localization depends on electrostatic interactions between the farnesyl moiety and anionic phospholipids [69], forming a farnesyl-electrostatic switch regulated by PKC [70] and Ca²⁺/calmodulin [71].

By no means being less necessary, S-prenylation has been shown to protect proteins against degradation due to enhanced folding and stability, a phenomenon described for Rab1B and YKT6 [72, 73]. Conversely, S-prenylation can also serve as degradation pathways for Rho proteins, thereby regulating protein turnover [74]. Returning to the subject, S-prenylation increases the affinity of proteins for interacting upstream or downstream signaling molecules, such as the interaction between KRas4B and hSOS1, which promotes GDP–GTP exchange and activates the Ras signaling cascade [75, 76].

Dual modification

These lipidations often act in concert (“dual lipidation”) to stably embed proteins at specific membrane microdomains. Notably, one lipid can “prime” the attachment of the second, for example, prenylation of Ras and N-myristoylation of Src-family kinases are prerequisite for their subsequent S-palmitoylation [77]. Src-family kinases (Lck, Fyn, Lyn, etc.), endothelial nitric oxide synthase (eNOS) [78,79,80,81], fibroblast growth factor receptor substrate (FRS2α) [82], LAMTOR1 [83], Gαi1 [13, 84, 85] are myristoylated at the N-terminus and then palmitoylated on nearby cysteines, a dual modification that tethers them to the lamellar structure of negatively charged phospholipids [84], lipid rafts [82], lysosome [83], or coated pits [80]. The sequential S-palmitoylation is evolutionarily encoded within the N-terminal pentapeptide (residues 2–6): proteins with Cys3 – 5 achieve ~ 75% N-myristoylation efficiency, of which 33–50% undergo S-palmitoylation due to additional Cys residues, reflecting a genetic imprinting strategy where weak myristoyl anchoring is reinforced by strong palmitoyl interactions [26] (Fig. 3) Table 1.

A parallel mechanism governs the dual modification of HRAS and NRAS, where S-prenylation is an obligatory first step that primes Ras for subsequent S-palmitoylation [86, 87]. Once palmitoylated, Ras is efficiently trafficked from the Golgi to the plasma membrane (PM), where it interacts with its regulators and effectors. Reversible removal of the palmitate by APT1/2 allows Ras to cycle back to the Golgi, maintaining a dynamic “on/off” localization loop. Kinetic studies have shown that the S-palmitoylation half-life of HRAS and NRAS ranges from approximately 30 min to 2 h [88], which is essential for fine-tuning their signaling outputs. In contrast, KRas4B lacks a S-palmitoylation site and instead relies on a polybasic tail, serving as an electrostatic membrane anchor [69], and the chaperone protein PDEδ to achieve stable PM association [89]. The presence or absence of S-palmitoylation also determines the specific membrane microdomains occupied by Ras isoforms. Palmitoylated HRAS and NRAS preferentially partition into order regions like lipid rafts and caveolae domains; but upon GTP loading it translocates out of rafts into the disordered (non-raft) regions of the PM, a shift critical for effective activation of downstream pathways like Raf/MEK/ERK [90]. If HRAS is artificially tethered in rafts continuously by adding an extra palmitate or preventing depalmitoylation, its ability to activate the Raf/MEK/ERK pathway is significantly impaired. Conversely, KRAS4B predominantly resides in non-raft regions regardless of its activation state, which affects its interaction with distinct effectors [90]. Thus, the dual lipidation of HRAS and NRAS provides a finely tuned mechanism for regulating signal amplitude and duration. Palmitoylation not only anchors Ras at the membrane to facilitate effective signal transduction but, owing to its reversible nature, also permits Ras to disengage, be recycled, or degraded—preventing overstimulation under normal physiological conditions.

By ensuring proper Ras localization and activation, dual lipidation facilitates crosstalk between Ras-driven oncogenic signals and the TME. Mutant Ras proteins are often locked in a sustained GTP-bound state, leading to continuous activation of the MAPK/ERK and PI3K-AKT pathways. This persistent signaling drives relentless cell proliferation, modulates immune responses by influencing the recruitment, activation, and differentiation of immune cells, and upregulates immune checkpoint molecules such as PD-L1 and CD47 [91].

Methodology for lipidation

Methods for studying protein lipidation face several challenges, primarily due to the low abundance of lipidated proteins in cells, the diversity of lipid types and modification sites, and the lack of specific labeling and enrichment strategies, which increase the difficulty of lipidation research.

Traditional methods

Traditional methods for studying lipidation, such as radioactive isotope labeling with compounds like [3H]palmitic acid [92], [3H]myristate, and [I125]myristate [13] have been valuable for detecting lipidation target proteins. Although this technique provided valuable tools for early research, it has limitations, including safety concerns, long detection times, and complex procedures. This method is usually unsuitable for large-scale screening due to its sensitivity and measurement throughput limitations. For the S-palmitoylation, alternative proteomics approaches, including acyl-biotin exchange (ABE) [93] and acyl-resin assisted capture (acyl-RAC) [94], exploit the labile nature of the thioester bond in S-palmitoylation to identify thousands of S-palmitoylation proteins through selective chemical cleavage with hydroxylamine. However, these methods have limitations, including high background and the inability to determine the acyl chain length [95].

Click chemistry

With recent advancements in chemical biology, click chemistry has emerged as a powerful approach for PTM studies, primarily due to its high specificity, efficiency, and compatibility with biological systems [96, 97]. Unlike traditional methods relying on radioactive labeling or antibodies, click chemistry uses lipid probes modified with alkynes or azide groups, which react selectively and efficiently with complementary fluorophores, affinity tags, or drug carriers via bioorthogonal cycloaddition reactions. This has enabled detailed analyses in diverse applications, including cell imaging, flow cytometry, gel electrophoresis, and immunoprecipitation [98]. A novel approach for studying N-myristoylation by designing photocrosslinkable and clickable myristic acid analog probes, enabling the capture of N-myristoylation-specific protein–protein interactions in living cells without genetic modification [99]. However, copper-catalyzed based click chemistry faces several challenges. The introduction of azides or alkynes and copper ions, can potentially alter the natural biological activity and raises concerns regarding potential toxicity [100]. To address these limitations, researchers have developed copper-free bioorthogonal reactions, such as strain-promoted azide-alkyne cycloaddition (SPAAC) [101] and tetrazine ligation [102]. The reaction kinetics of SPAAC is slower than that of CuAAC, but its biocompatibility in living cells is unquestionable [103]. Tetrazine ligation, involving inverse electron-demand Diels–Alder (IEDDA) reactions between tetrazines and strained alkenes or alkynes, offers rapid reaction rates and excellent bioorthogonality, making it particularly suitable for in vivo applications [102]. However, these bioorthogonal cycloaddition reactions often use tensile olefins or alkynes as reactants, which are difficult to synthesize and have poor stability. Additionally, novel bioorthogonal cycloaddition reactions have been developed to simultaneously study multiple biomolecules in complex systems, providing new strategies for investigating lipidation processes [104].

Mass spectrometry

Liquid chromatography-tandem mass spectrometry (LC–MS/MS) and matrix-assisted laser desorption/ionization mass spectrometry imaging (MALDI-MSI) are pivotal techniques in the study of lipidation. LC–MS/MS integrates liquid chromatography with mass spectrometry to separate peptides and analyze their sequences, enabling the identification and quantification of lipidation [105]. Since lipidation is a low-abundance modification, enrichment strategies are essential for detection. Common approaches include click chemistry-based metabolic labeling, genetic encoding of unnatural amino acids (UAAs) with in vivo installation of palmitate mimics via the IEDDA reaction [106], ABE [107], and the Acyl-RAC all of which improve sensitivity in LC–MS/MS analysis and facilitate dynamic quantification of S-palmitoylation [108]. This method is highly sensitive and specific, making it indispensable in proteomics research. For example, LC–MS/MS has been employed to profile human protein S-palmitoylation [109], and our lab utilizes this technique to identify substrates of ZDHHC20 and map S-palmitoylation sites [110]. Notably, the click chemistry proteomics strategy takes advantage of the unique reversibility of S-palmitoylation, combined with pulse-chase methods to observe the reversible changes of protein modifications over time, revealing their dynamic regulatory mechanisms [111].

Conversely, MALDI-MSI allows for the in situ analysis of proteins and their modifications within tissue sections, preserving spatial context. By ionizing molecules directly from tissue surfaces, MALDI-MSI provides detailed maps of protein distributions and their PTMs across different tissue regions [112]. This technique has been applied to enhance the detection of extracellular matrix (ECM) proteins, which are challenging to analyze due to their size, insolubility, and extensive modifications like glycosylation. Innovative methods combining tissue decellularization with enzymatic digestion have improved ECM protein detection using MALDI-MSI, facilitating more accurate spatial mapping of these proteins [113]. For palmitoylated proteins, a major challenge is detecting the intact protein or peptide. A clever solution is on-tissue chemical derivatization [114]; for example, treating a tissue section with hydroxylamine cleaves thioester-linked palmitate, producing a predictable mass shift, which MALDI-MSI can then detect to generate a spatial map of palmitoylated protein abundance. MALDI-MSI can achieve a spatial resolution of 5–10 μm [115], but since it lacks chromatographic separation, coexisting biomolecules such as proteins, lipids, and inorganic salts can interfere with desorption and ionization efficiency. Therefore, optimized tissue preparation methods are required to improve detection accuracy [116]. LC–MS/MS excels in detailed characterization and quantification of PTMs in complex protein mixtures, while MALDI-MSI provides valuable insights into the spatial distribution of modified proteins within tissues.

Recent advancements in lipidation research have leveraged metabolic labeling, click chemistry, and nanoLC-MS/MS-based proteomics to achieve high-throughput, site-specific analysis of protein S-prenylation [117], this study introduces a novel approach utilizing alkyne-tagged isoprenoid analogs (YnF and YnGG) for bioorthogonal labeling of prenylated proteins in living cells, followed by selective enrichment via click chemistry. This strategy overcomes longstanding challenges in detecting lipid modifications, including hydrophobicity and poor ionization efficiency, which have historically hindered mass spectrometric analysis. The method demonstrates high sensitivity, as evidenced by the low nanomolar to micromolar IC₅₀ values observed for FTase inhibitors, highlighting its effectiveness in quantifying lipidation dynamics with precision. Furthermore, advanced quantification strategies such as stable isotope labeling by amino acids in cell culture (SILAC) [118, 119] or tandem mass tagging [120] improve accuracy and reproducibility.

Advanced imaging technologies

Unlike click chemistry, imaging techniques such as Förster Resonance Energy Transfer (FRET) allow real-time observation of molecular interactions within living cells, providing continuous, non-invasive monitoring of dynamic biological processes. While click chemistry primarily detects biomolecular events at specific time points, offering "snapshots" reflecting target molecule concentrations (with detection limits as low as 1.8–7.3 μg L⁻¹[121], FRET focuses on the spatial dimension, detecting interactions within the 1–10 nm range [122]. Thus, FRET is highly sensitive for studying protein–protein interactions and visualizing biological processes such as T cell signals and the signal transduction [123] in the TME. Recently, FRET has been increasingly utilized in lipidation studies to investigate the distances and interactions among lipidated proteins by labeling them with donor and acceptor fluorophores [124]. One effective approach involves engineering reporter proteins that exhibit FRET signal changes upon lipidation. For instance, a fluorescent protein fused with a known lipidation-target sequence translocates to cellular membranes when lipidated, changing the proximity between donor and acceptor fluorophores [125, 126]. This allows dynamic, real-time observation of lipidation events within living cells, providing high sensitivity capable of detecting subtle structural and functional changes in proteins. Additionally, FRET spectroscopy can analyze fluorescence variations to infer the size, shape, and composition of protein complexes, offering insights comparable to mass spectrometry [127]. However, this technology still faces challenges, such as the lengthy exploration of fluorescence donors and acceptors, as well as the need to carefully consider the spectral overlap between the donor and acceptor, which demands high experimental precision.

With the continuous advancement of imaging technologies, fluorescently labeled multifunctional peptides have also been used to study lipid-modified proteins. For example, multifunctional peptides labeled with fluorescent tags based on the C-terminus of naturally isoprenylated protein cell division control protein (CDC42) [128] or by introducing ketone into myristoylated ADP-ribosylation factor 1 (ARF1) proteins followed by fluorescence labeling with hydrazine conjugated to aminobenzene azides, allow the modified proteins to be detected by fluorescence microscopy [129], enabling localization and dynamic observation of these proteins within living cells.

Moreover, super-resolution microscopy has provided a new perspective on the visualization of lipid modifications. For instance, combining a novel bioorthogonal cholesterol probe (chol-N₃) with super-resolution microscopy allows for direct visualization and characterization of lipid rafts in living cells with unprecedented resolution [130]. This approach, which combines bioorthogonal chemistry and super-resolution microscopy, offers new possibilities for dynamic lipidation observations, and it is expected that this method will be widely applied in lipid modification research in the future. Significant advances have recently been made in super-resolution stimulated Raman scattering (SRS) microscopy. The development of the Adam point-cloud deconvolution algorithm has improved the spatial resolution of SRS microscopy, enabling the reconstruction of high-resolution images from SRS data using deep learning techniques. This makes it possible to conduct nanoscale label-free imaging and study metabolic dynamics in living cells [131], which provides a more precise and in-depth tool for studying the dynamic processes of lipidation.

Multi-omics integration

With more profound studies of multi-omics and the development of multi-dimensional data such as spatial transcriptomics, researchers have been able to delve deeper into the dynamics of cells in space and time [132]. The growing recognition of biological interconnectedness has driven the integration of various omics methods (e.g., proteomics, metabolomics, and lipidomics) to provide a more comprehensive view of lipidation networks. Informatics tools have significantly supported this integration, with platforms like LipidSig [133], LIPID MAPS [134] provide comprehensive, continuously updated computational pipelines and databases tailored specifically for the lipidomics research communities. Moreover, the NCI’s Clinical Proteomic Tumor Analysis Consortium (CPTAC) has integrated mass-spectrometry-based proteomic data with lipidomics and metabolomics data across various cancers, including breast, renal, and colon cancers, thereby enabling detailed correlations between lipid profiles, protein expressions, and genomic alterations [135]. Multi-omics approaches have been instrumental in elucidating the functional roles of lipidation. For example, studies combining differential proteomics and transcriptomics have shown that N-myristoylation regulates mitochondrial respiratory complex I, thereby impacting cellular energy metabolism and signaling [136]. Another study integrated genomics, transcriptomics, immunomics, and proteomics, offering strong data support for understanding the role of S-palmitoylation in cancer [137]. As multi-omics integration becomes a trend in lipidation research, researchers can study lipidation processes more comprehensively, enhancing our understanding of its role in cellular functions and disease. Moreover, multi-omics data integration has provided a clearer view of the complex interactions between lipidation and other post-translational modifications, such as lipidation with phosphorylation and ubiquitination, which are critical for tumor signaling pathways. For example, studies show that ZDHHC3/9 modulates the tumor immune microenvironment by inhibiting PD- 1/L1 ubiquitination [21, 138], while our lab has found that ZDHHC20 stabilizes FASN by preventing its ubiquitin–proteasome mediated degradation, promoting liver cancer progression [110]. Phosphorylation also plays a role in the S-palmitoylation of NLRP3, regulating inflammasome activation and impacting immune responses and tumorigenesis [139].

Artificial intelligence

As multi-omics data continues to grow in scale and complexity, managing, integrating, and analyzing this information pose significant challenges. Artificial Intelligence (AI) in medicine plays a critical role in addressing these challenges by enabling multi-layer data integration, predictive modeling, and drug discovery [140]. One such advancement is multiDGD, a deep generative model that processes transcriptomic and epigenomic data, detects statistical associations between genes and regulatory regions, and reveals the relationship between gene expression and epigenetic regulation [141]. AI has proven particularly valuable in lipidation research, allowing for the prediction of modification sites and the discovery of novel inhibitors. For instance, ISM6331, identified through AI-generated models, was found to be a potent TEAD S-palmitoylation inhibitor, demonstrating reversible binding to TEAD S-palmitoylation sites and strong suppression of TEAD transcriptional activity [142]. Notably, ISM6331 has received orphan drug designation and entered Phase I clinical trials (NCT06566079), with the first patient dosed in January 2025.

Additionally, AI-based tools like SwissPalm2 [143] and deep learning-based graphical models [144] have been developed to predict S-palmitoylation sites in proteins, expanding computational capabilities in lipidation studies. AI has also advanced research on N-myristoylation and S-prenylation. A recent study leveraged machine learning models—including extreme learning machines, incremental feature selection, and maximum relevance minimum redundancy —to develop an optimal N-myristoylation site prediction model, enhancing the identification, understanding, and large-scale analysis of N-myristoylation in protein sequences [145]. Similarly, a machine learning-based approach to enhance S-prenylation site prediction, analyzing both canonical (prenylated, cleaved, and carboxymethylated) and shunted (prenylation-only) motifs, covering 8,000 possible Cxxx sequence combinations [146].

AI is also transforming cancer metabolism research by identifying metabolic enzyme targets that influence both tumor growth and immunity. Using BipotentR, an AI-based multi-omics analysis tool, researchers identified ESRRA (estrogen-related receptor α) as a key regulator of tumor metabolism and immune suppression. Inhibiting ESRRA effectively blocked lipid oxidation in tumor cells, leading to cancer cell death while simultaneously enhancing anti-tumor T cell responses in preclinical models [147]. AI technology is also advancing in drug discovery [148] and bioinformatics, enabling the identification of disease biomarkers [149] and optimizing lipid nanoparticle formulations to improve drug delivery efficiency [150]. These advancements demonstrate that AI holds great potential for lipidation-related disease treatment by enhancing the quantitative analysis, dynamic tracking, and drug development of lipidation.

Roles of lipidation in the TME

Tumor cell survival and death signaling

The TME is a complex ecosystem where tumor cells must balance proliferation, metabolic adaptation, and resistance to various forms of cell death to sustain malignancy [151]. Lipidation multiple facets of tumor biology, including signaling, metabolism, and stress adaptation, allowing cancer cells to evade apoptosis, ferroptosis, and other cell death mechanisms [152]. By fine-tuning protein localization, stability, and function, lipidation plays a central role in determining tumor cell fate [4].

Tumor cell proliferation and signal transduction

Uncontrolled proliferation is a defining feature of cancer, supported by dysregulated oncogenic signaling and disrupted cell cycle checkpoints [153]. Lipidation modifies key signaling molecules that drive cell division, growth factor receptor activation, and mitotic progression.

NMT1, which is upregulated in various cancers, including oral squamous cell carcinoma (OSCC) [154], liver cancer [155], lung cancer [19], breast cancer (BrCa) [156], osteosarcoma [157], and bladder cancer [158], is frequently associated with poor prognosis [159]. As one of the earliest identified myristoylated proteins, Src is crucial in mediating ECM-integrin interactions, linking the ECM to the cytoskeleton, and inducing cell adhesion turnover [160]. In the TME, Src N-myristoylation, particularly when enhanced by exogenous myristic acid, promotes the activation of the FAK-Src signaling axis [161] (Fig. 4). This leads to an amplification of FAK phosphorylation and yes-associated protein (YAP)/transcriptional coactivator with PDZ-binding motif (TAZ)-dependent IL- 6 expression, reinforcing the tumor-stroma crosstalk. Src N-myristoylation also facilitates the secretion of Periostin by fibroblasts, contributing to ECM remodeling and sustaining the TME homeostasis through an IL- 6/STAT3 feedback loop, which ultimately promotes colorectal cancer progression [162]. The mitogen-activated protein kinase (MAPK)/ERK pathway, a key regulator of tumor proliferation, is initiated when receptor tyrosine kinases (RTKs), such as epidermal growth factor receptor (EGFR) or fibroblast growth factor receptor (FGFR), activate RAS GTPases. Upon activation, RAS localize to the PM to engage RAF kinases, triggering ERK phosphorylation and subsequent transcription of cell cycle-promoting genes [163]. The membrane anchoring of RAS is controlled by S-prenylation and S-palmitoylation (Fig. 6). Farnesylation at the CAAX motif facilitates the initial recruitment of RAS to the PM, but stable retention of HRAS and NRAS requires S-palmitoylation [86, 87]. They translocate to the membrane and initiate the Ras-Raf-MEK-ERK pathway [164, 165]. The absence of S-prenylation in certain RAS-related inhibitors prevents their membrane localization, thereby impairing their pro-tumorigenic effects [166].

Lipidation also regulates the phosphoinositide 3-kinase (PI3 K)/AKT/mTOR pathway, which controls cell growth, survival, and metabolic adaptation [167]. KRAS-dependent activation of PI3 K relies on its S-palmitoylation, which enhances its interaction with PI3 K and promotes AKT activation. ZDHHC13 [168] and ZDHHC20 [169, 171] mediated S-palmitoylation sustain PI3 K/AKT signaling through EGFR, driving uncontrolled proliferation. ABHD17 C affects the PI3 K/AKT signaling pathway by causeing cell cycle arrest, and reducing cell proliferation [172]. Downstream of AKT, mTORC1 activation is controlled by Rheb, whose S-prenylation is required for its lysosomal localization and function [172, 173]. When PDEδ, a S-prenylation chaperone, is inhibited, Rheb mislocalizes, leading to defective mTORC1 activation and impaired tumor growth [174]. Disruption of upstream and downstream events in S-prenylation, such as inhibition of the mevalonate (MVA) pathway or PDEδ, leads to the arrest of growth and induced cell death [164, 165]. Additionally, S-palmitoylation differentially affects mTOR signaling depending on cancer type: in renal cell carcinoma, ZDHHC2 enhances AGK S-palmitoylation, leading to AKT-mTORC1 activation [175], whereas in BrCa, ZDHHC22 reduces mTOR stability, thereby suppressing AKT-mTOR signaling [176].

The Hippo pathway is a highly conserved regulatory network regulates cell proliferation by YAP/TAZ-TEA domain family member (TEAD) transcriptional activity [177]. TEAD auto-palmitoylation is essential for its stability and interaction with YAP/TAZ, driving proliferative gene expression [178, 179]. Given its role in sustaining uncontrolled growth, inhibiting TEAD S-palmitoylation has emerged as a promising strategy for blocking Hippo pathway activation in cancers.

Additionally, N-myristoylation promotes or stabilizes liquid–liquid phase separation (LLPS), imparting hydrophobicity to proteins and increasing their signal transduction efficiency [180, 181]. In lung cancer, the N-myristoylation of enhancer of zeste homolog 2 (EZH2) facilitates the formation of LLPS droplets that concentrate the substrate signal transducer and transcription 3 (STAT3) activator, thereby increasing STAT3 signaling and driving cell proliferation [180](Fig. 4).

Interestingly, the effects of lipidation on tumor progression vary depending on the tissue type, developmental stage, and signaling environment. For instance, S-prenylation-related enzyme deletion can have paradoxical outcomes: inactivation of Rce1 can promote myeloproliferative diseases [182], whereas conditional ablation of ICMT may impair oncogenicity mediated by all RAS isoforms [183]. In some instances, concurrent deactivation of KRAS G12D and ICMT may accelerate disease progression in embryonic pancreatic cancer models [184]. Different conclusions hinge on the tissue type, stage of development, and signaling environment.

Lipidation also plays a role in cell cycle regulation and mitotic checkpoint control. Cell cycle regulation determines whether cells continue to proliferate or enter cell death pathways. NMT inhibitors induce cell cycle arrest by blocking the PI3 K and MAPK signaling pathways [185, 186], whereas the myristylation of sphingosine kinase 1 delays exit from the G₀/G₁ phase [17]. Additionally, centromere proteins CENP-E and CENP-F, expressed during mitosis to mediate the G2 to M checkpoint, are regulated by S-prenylation, which modulates their interaction with the RZZ complex and ensures proper chromosome segregation while inhibiting farnesylation of the dynein adaptor Spindly causes mitotic defects and results in metaphase arrest, preventing chromosome separation [187,188,189]. Blocking S-prenylation upregulates the expression of the cyclin-dependent kinase (CDK) inhibitors p21 and p27, inhibits the activity of CDK2 and the expression of cyclins D1 and E2, and blocks the MAPK/ERK and PI3 K/Akt pathways, thus halting cell cycle progression [190,191,192,193,194,195]. This inhibition also increases DNA damage and induces cellular senescence [196].

Tumor metabolism and energy adaptation

Cancer cells undergo metabolic reprogramming to meet their high demands for adenosine triphosphate (ATP), lipids, and metabolic intermediates, enabling them to adapt to metabolic stress [197]. Some mitochondrial proteins require N-myristoylation for proper mitochondrial targeting and respiratory chain assembly [198, 199], ensuring efficient oxidative phosphorylation (OXPHOS) and ATP production [136]. For instance, NDUFB7, an accessory subunit of mitochondrial complex I, and DMAC1, a critical assembly factor, require N-myristoylation for their mitochondrial targeting and function [15]. N-myristoylation also regulates the mitochondrial localization of proteins like Cyb5R3 and DES1, influencing sphingolipid metabolism and altering cancer cell metabolism [200,201,202]. Similarly, N-myristoylation at the N-terminus of Mic19 enhances its interaction with the mitochondrial outer membrane and Tom20, facilitating its efficient import into the mitochondrial intermembrane space [203]. Inhibition of NMT1 leads to mitochondrial dysfunction and energy crisis, ultimately leading to cancer cell death. Further, the N-myristoylation of AMPK enables metabolic plasticity, supporting the transition from OXPHOS to glycolysis. Inhibition of AMPK N-myristoylation prevents the metabolic adaptation of cancer cells, making them more susceptible to cell death [204, 205].

S-palmitoylation is crucial in regulating tumor cell glycolysis and metabolic adaptation, particularly its impact on lactate dehydrogenase A (LDHA) [16] (Fig. 5). LDHA catalyzes the conversion of pyruvate to lactate, a critical step in glycolysis, especially under hypoxic conditions [206]. LDHA levels are elevated in tumors, mainly due to their role in driving lactate production while bypassing mitochondrial OXPHOS. This metabolic shift enables cancer cells to thrive in hypoxic environments, promoting tumor growth and invasion [207]. Increased lactate output further acidifies the TME, supporting tumor invasion and metastasis [208]. Additionally, S-palmitoylation indirectly affects the production of ROS by modulating glycolysis. The shift from mitochondrial OXPHOS to glycolysis reduces ROS generation, producing less ROS than mitochondrial respiration [209]. Thus, the S-palmitoylation of LDHA for tumor cells minimizes oxidative stress, allowing cell survival and proliferation, especially under metabolic or hypoxic stress conditions [206]. MDH2 is a key enzyme in the TCA cycle, activated by S-palmitoylation at position C138. This modification enhances the catalytic activity of MDH2, promotes ATP production, and supports tumor cell proliferation [210]. Therefore, S-palmitoylation is a "metabolic switch" that helps cancer cells survive and proliferate in hypoxic or nutrient-poor environments. Therefore, inhibition of the specific PATs of MDH2 or LDHA may be a promising therapeutic strategy to target tumors. However, studies have shown that depalmitoylated KRAS4 A promotes glycolysis through its interaction with hexokinase 1, thereby meeting the increased energy demands required for tumor proliferation. This process is closely correlated with the copy number of mutant KRAS, independent of the MAPK signaling pathway [211]. Regarding lipid metabolism, S-palmitoylation regulates endothelial lipase and the peroxisome proliferator-activated receptor gamma–ATP citrate lyase axis through S-palmitoylation [212, 213]. Notably, the downregulation of acyl-CoA oxidase 1 modulates the accumulation of palmitic acid (PA), subsequently affecting the S-palmitoylation of β-catenin and influencing tumor progression [214].

The S-prenylation-related process affects tumor metabolism. TCS Blockade of the MVA pathway leads to reduced levels of coenzyme Q10, a critical electron transport protein in the mitochondrial respiratory chain, and polyisoprenoids, vital free radical scavengers in the cell membrane. This disruption causes a decrease in mitochondrial membrane potential, promotes the release of pro-apoptotic factors, and induces mitochondrial dysfunction, thus exacerbating oxidative stress And inhibiting Rab protein S-prenylation can induce the unfolded protein response and disrupt purine metabolism [215]. Inhibition of ICMT severely affects mitochondrial function and inhibits ATP production by affecting electron transport chains I, II, and III. Inadequate energy supply leads to a decrease in key metabolites in the TCA, inhibiting anabolic processes and cell growth. Particularly in ICMT-inhibitor-sensitive pancreatic cancer cells, ICMT inhibition causes mitochondrial respiratory defects and energy depletion, leading to a significant upregulation of p21. p21 serves as a sensor of cellular energy depletion and activates BNIP3 to trigger the apoptotic response [216, 217].

Tumor cell stress responses and cell death regulation

Autophagy

Autophagy is a vital cellular process that maintains homeostasis in cancer cells through lysosomal degradation mechanisms. It plays a dual role in cancer, acting both as a tumor suppressor by clearing damaged organelles and as a tumor promoter by sustaining metabolic adaptation and survival under stress [218].

Autophagy is tightly linked to mechanistic target of rapamycin complex 1 (mTORC1), a central regulator of nutrient sensing and cellular metabolism [219]. N-myristoylation and S-palmitoylation fine-tune the balance between mTORC1 activity and autophagy initiation. N-myristoylation and S-palmitoylation stabilizes lysosomal recruitment of LAMTOR1, a key component of the Ragulator complex, which activates mTORC1 under nutrient-rich conditions [158, 220](Fig. 4). Inhibiting N-myristoylation of LAMTOR1 disrupts mTORC1 localization, impairing its ability to suppress autophagy [83]. AMPK N-myristoylation enhances mitophagy [221, 222], while ARF1 N-myristoylation activates stimulator of interferon (IFN) gene (STING)-dependent autophagy [223], contributing to metabolic adaptation in cancer cells. Conversely, N-myristoylation of protein kinase B (AKT) suppresses autophagy by reducing Beclin- 1, LC3a, and LC3b expression while increasing p62 levels, thereby shifting the cell towards survival instead of degradation [224]. In BrCa, blocking NMT1 increases protein degradation and ER stress, triggers the c-Jun N-terminal kinase (JNK) pathway, and stimulates autophagy, thus contributing to tumorigenesis [156].

Autophagy initiation requires the unc- 51-like kinase (ULK) and class III PI3 K complexes [225], with ULK1 acting as a central regulatory hub. Upon nutrient deprivation or mammalian target of rapamycin (mTOR) inhibition, ULK1 activation is triggered through dual regulatory mechanisms: AMPK-mediated phosphorylation (Ser317/Ser777) directly enhances its kinase activity [226], while ZDHHC13-dependent S-palmitoylation facilitates ULK1 complex translocation to the autophagosome formation site [227] (Fig. 5). This modification not only complements AMPK-driven phosphorylation but also potentiates ULK1's ability to phosphorylate ATG14L within the VPS34 complex, thereby amplifying PI3 K lipid kinase activity – a mechanism evolutionarily conserved from C. elegans to mammals [227, 228]. Intriguingly, ZDHHC17 further modulates this network by priming AMPK activation [229]. Creating a feedforward loop that ensures robust autophagy initiation. Recent studies challenge the linear view of AMPK as a simple "on-switch", revealing its context-dependent roles: while sustaining basal autophagy, AMPK paradoxically attenuates acute autophagic responses under prolonged energy stress [230]. Moreover, DHHC5-mediated S-palmitoylation of Beclin 1 enhances its interaction with ATG14L and VPS15, further promoting autophagy [231]. After the initial stages of autophagy, S-palmitoylation promotes autophagosome formation by modifying ATG16L1 [232,233,234]. This modification promotes the interaction of ATG16L1 with WIPI2B and Rab33B, thereby ensuring efficient autophagosome generation [233]. Even calcium homeostasis during autophagy is regulated via S-palmitoylation of MCOLN3/TRPML3, a Ca²⁺-permeable cation channel [235]. On the other hand, PPT1 deficiency disrupts the removal of S-palmitoylation, affecting inositol 1,4,5-trisphosphate receptor 1 expression and thereby dysregulating lysosomal Ca homeostasis [236]. Moreover, PPT1 deficiency disrupts Rab7 localization and impairs autolysosome function [237]. The selective autophagy receptor p62 and autophagy substrates are also regulated by S-palmitoylation, further refining the selectivity and efficiency of the autophagic process. Specifically, p62 is palmitoylated by ZDHHC19 to increase its affinity for LC3-positive autophagosomes, thus strengthening the selective degradation of ubiquitinated proteins [238]. In contrast, ZDHHC5 prevents NOD2 P62-mediated degradation by maintaining its membrane localization [239].

An S-prenylation-related growth-promoting protein PRL- 3 drives the PIK3 C3-Beclin- 1-dependent autophagosome formation to accelerate cell autophagy [240]. Autophagy is negatively regulated by the ICMT [241]. ICMT inhibition induces autophagy and acts as an upstream event of apoptosis in some tumors. This suggests that targeting ICMT can activate a cascade of cellular events that promote cell death. The synergistic effect of both processes can be achieved via ICMT inhibitor cysmethynil, resulting in substantial antitumor effects [242]. ICMT inhibition also suppresses mitochondrial respiration, leading to cellular energy depletion and increased autophagy, thus providing further insight into how ICMT inhibition contributes to autophagy regulation [216].

Apoptosis

Apoptosis is a programmed cell death mechanism that eliminates damaged or stressed cells without triggering inflammation. This process is tightly regulated by caspase activation, cytochrome C release, and pro-apoptotic signaling cascades [243]. Apoptosis is triggered through caspase-mediated proteolytic cleavage, predominantly via glycine residues, to eliminate damaged cells with the advantage of not inducing an inflammatory response [244]. Posttranslational N-myristoylation occurs when caspases cleave proteins, exposing an internal glycine residue that serves as a lipidation site (Fig. 4). Modifying the resulting C-terminal fragment is crucial for balancing cell survival and apoptotic pathways [245,246,247]. N-myristoylation plays a dual role in apoptosis, influencing both pro-apoptotic and anti-apoptotic pathways. N-myristoylation of proapoptotic proteins, including C-terminal BID (ctBID) [248] and C-terminal p21-activated protein kinase 2 (ctPKA2) [249], targets them to mitochondrial membranes and cristae, increasing cytochrome C release and activating the JNK pathway, thereby promoting apoptosis. Conversely, the N-myristoylation of antiapoptotic proteins, such as ctGelsolin [250] and ctPKC [251], inhibits apoptosis [252]. N-myristoylation also regulates apoptosis through apoptosis-related proteins. For example, ARF1 N-myristoylation negatively regulates apoptosis by activating RSK1 and phosphorylating BAD [253, 254]. N-myristoylation of FUS1 induces apoptosis by activating Bcl- 2-associated X (BAX)and caspases [255]. In addition, the N-myristoylation of DES1 targets it to the mitochondria, affecting mitochondrial sphingolipid metabolism, particularly ceramide production. Ceramide facilitates cytochrome C release and disrupts the electron transport chain [200, 256, 257]. Interestingly, caspases regulate the cleavage and relocalization of NMTs. In cancer cells with diminished caspase activity, NMT cleavage is impaired, leading to persistently high NMT expression, which supports anti-apoptotic signaling [258]. Thus, NMT inhibition reduces Bcl- 2 expression and increases cleaved caspase- 3 levels, inducing apoptosis in tumor cells [158, 259].

S-palmitoylation modulates apoptotic signaling and thus determines cell fate under conditions of stress [260]. In the extrinsic apoptosis, it enhances apoptotic signal transduction by regulating the membrane localization and oligomerization of TNF superfamily members, such as Fas and DR4 [39, 261, 262]. DR4 in lipid rafts increases the sensitivity of rectal cancer cells to TRAIL-mediated apoptosis [263, 264]. Similarly, S-palmitoylation increases Fas-mediated caspase- 8 activation and promotes apoptosis [262, 265]. In endogenous apoptosis, BAX S-palmitoylation is necessary for binding to the mitochondrial outer membrane, whereas nonpalmitoylated BAX reduces its apoptotic activity [266]. Moreover, ZDHHC9 palmitoylates immunoglobulin protein, which enhances cell tolerance to ER stress, consequently reducing apoptosis [267]. Consequently, the knockout of ZDHHC9 induces apoptosis in lung and bladder cancer tumor cells [267, 268]. In BrCa, S-palmitoylation of neurotensin receptor- 1 enhances its interaction with Gαq/11, thus inhibiting apoptosis [269].

Blocking the MVA pathway inhibits S-prenylation, reducing the availability of FPP and GGPP, thereby preventing protein S-prenylation and altering apoptotic sensitivity [270, 271]. S-prenylation of Rab family proteins (e.g., Rab7) regulates vesicle trafficking. In drug-resistant cancer cells, disrupting Rab7 S-prenylation leads to excessive ROS accumulation, which ultimately triggers apoptosis [215, 272]. MVA pathway inhibition simultaneously activates intrinsic apoptotic pathways by activating Bcl- 2, Caspase- 3, PARP, and c-Jun, leading to cancer cell apoptosis [273, 274]. In KRAS-mutant cells, S-prenylation inhibition also results in severe ER stress [275]. Deprenylation triggers the extrinsic apoptosis pathway by increasing DR5 expression, enhancing TRAIL-mediated apoptotic signaling, activating caspase- 3, and promoting tumor cell death [276].

Ferroptosis

Ferroptosis is a nonapoptotic form of cell death characterized by iron-dependent lipid peroxidation [277]. This process involves iron accumulation, lipid peroxidation, ROS production, and depletion of glutathione peroxidase 4 (GPX4) [278]. Owing to their increased iron requirements, cancer cells are particularly vulnerable to ferroptosis, making its regulation a critical determinant of tumor survival [279, 280].

N-myristoylation critically regulates ferroptosis by targeting ferroptosis suppressor protein 1 (FSP1) to the PM, increasing the reduction in coenzyme Q₁₀(CoQ10) and reducing the accumulation of lipophilic radicals and lipid peroxides [281]. Mutations at N-myristoylation sites on FSP1 impair membrane localization and ferroptosis inhibitory function [282](Fig. 4). In addition, ACSL1-mediated FSP1 N-myristoylation increases resistance to ferroptosis in ovarian cancer [283]. N-myristoylation also enhances the interaction between icFSP (an FSP1 inhibitor) and FSP1, promoting FSP1 phase separation and impairing tumor growth [284]. N-myristoylation inhibition triggers extensive lipid peroxidation and ferroptosis and activates the parthanatos pathway, increasing tumor cell sensitivity to ferroptosis-induced necrotic apoptosis [18, 19].

SLC7 A11 has a crucial role in maintaining cellular antioxidant defense and redox homeostasis by importing cystine for SLC7 A11 glutathione synthesis and regulates ferroptosis by preventing oxidative damage and lipid peroxidation [285]. ZDHHC8 stabilizes SLC7 A11, an essential regulator of ferroptosis, by decreasing its ubiquitination and degradation. This can be further enhanced by the long-chain noncoding RNAs DUXAP8 and AMPK, which inhibit ferroptosis [286, 287]. Moreover, PPT1 upregulates the expression of the antioxidant enzyme GPX4, further inhibiting ferroptosis in OSCC [288] (Fig. 5).

S-prenylation regulates ferroptosis primarily through the MVA pathway by modulating the maturation of selenocysteine tRNA, thus affecting GPX4 synthesis and FSP1 activity [289, 290] and influencing ferroptosis occurrence. Products of the MVA pathway, such as isopentenyl pyrophosphate and CoQ10, inhibit lipid peroxidation; therefore, their depletion is believed to induce ferroptosis [289,290,291].

Pyroptosis and inflammasome activation

Pyroptosis is an inflammation-driven, caspase- 1-dependent, and gasdermin-mediated form of programmed cell death. It is characterized by cell swelling and the formation of bubble-like protrusions on the plasma membrane before rupture, leading to the release of inflammatory factors. This process is mediated by inflammasome assembly, which triggers the immune response [292].

Pyroptosis depends on gasdermin D (GSDMD) family proteins, and ZDHHC5/7/9-mediated S-palmitoylation of GSDMD at the C191 site facilitates its activation and translocation to the cell membrane after NLRP3 inflammasome cleavage, where it then forms pores and triggers pyroptosis [293]. Whereas prior studies theorized that GSDMD must undergo cleavage to produce the GSDMD N-terminal domain (GSDMD-NT) and create large transmembrane pores, a more recent study challenged this notion by showing that reversible S-palmitoylation is an essential checkpoint for pore formation by both GSDMD-NT and intact GSDMD, [294] (Fig. 5). The multiple roles of S-palmitoylation show that the depalmitoylating enzyme APT2 promotes the activation of GSDMD by removing palmitoyl groups. Nevertheless, the knockout of ABHD17 C enhances pyroptosis responses in liver cancer cells [295], highlighting the crucial balance between S-palmitoylation and depalmitoylation in pyroptosis [293]. Moreover, a S-palmitoylation-related GSDMD inhibitor, NU6300, was invented and has an elite effect [296].

The NLRP3 inflammasome comprises NLRP3, apoptosis-associated speck-like protein (ASC), and pro–caspase–1, which collectively facilitate the cleavage and activation of GSDMD [297] (Fig. 4). S-palmitoylation represents a crucial regulatory switch in NLRP3 inflammasome activity. Indeed, ZDHHC1 enhances the interaction between NLRP3 and NEK7, promoting inflammasome assembly [298]. ZDHHC5 of NLRP3 increases its degree of membrane binding and stability, further supporting practical inflammasome function. In contrast, ABHD17 A and ZDHHC12 act to prevent hyperactivated inflammatory responses by depalmitoylating or inhibiting NLRP3, thereby preventing overactivation of the inflammasome [299, 300]. Similarly, inhibiting fatty acid synthase (FASN) blocks the S-palmitoylation of NLRP3, resulting in reduced inflammasome activation [301].

The geranylgeranylation of RhoA is essential for proper NLRP3 inflammasome assembly [302] ensuring effective GSDMD-dependent pyroptotic signaling [303]. Additionally, reduced S-prenylation of Rac1 enhances its interaction with IQGAP1, further stimulating NLRP3 inflammasome activation and promoting pyroptosis [304].

Cell plasticity and dynamic behavior

Epithelial-mesenchymal transition (EMT) and MET are no longer viewed as binary processes but as dynamic transitions where cells can adopt hybrid states with both epithelial and mesenchymal traits, exhibiting epithelial-mesenchymal plasticity [305]. This plasticity allows tumor cells to adapt and respond to various microenvironmental cues, increasing their capacity for invasion and metastasis.

For instance, N-myristoylation plays a crucial role in EMT by modulating the FGF/FGFR signaling axis. The transition between FGFR2 IIIb and FGFR2 IIIc isoforms, which is key to mesenchymal transformation, is influenced by the AKT3/IWS1 pathway [306, 307]. The scaffold protein FRS2α, a key mediator of FGF signaling, undergoes N-myristoylation, which stabilizes FGFR/Src interactions, amplifying downstream signaling and promoting EMT plasticity [186](Fig. 4). Concurrently, S-palmitoylation of transcriptional regulators such as Snail reprograms lipidation networks within EMT by activating APT2 [308], a depalmitoylating enzyme that displaces Scribble from the PM, disrupting Hippo-YAP signaling and enhancing the MAPK pathway—key drivers of mesenchymal identity [309] (Fig. 4).

At the cytoskeletal level, S-prenylation interlocks with these lipid modifications to orchestrate cytoskeletal remodeling. The S-prenylation-dependent activation of RhoA and Rac1, facilitated by ICMT [310], reinforces actomyosin contractility and cell shape changes required for EMT progression and tumor metastasis [311]. In particular, TP53 mutations hyperactivate RhoA S-prenylation through the MVA pathway, amplifying RhoA/ROCK1/actomyosin mechano-signaling, which suppresses Lats1/2 activity and enhances TEAD/YAP transcriptional activity, reinforcing a metastatic phenotype [312,313,314] (Fig. 6). Beyond EMT regulation, lipidation is a central orchestrator of tumor cell movement, determining adhesion, cytoskeletal flexibility, and ECM interactions. N-myristoylation of MARCKS enhances fibronectin adhesion, stabilizing cell–matrix interactions [315, 316], yet paradoxically, prenylated RhoA activation simultaneously primes actin polymerization, fostering detachment and migration [313, 317]. This lipidation-driven tug-of-war enables tumor cells to finely regulate adhesion strength, balancing cell–matrix interactions with the need for invasive mobility.

At the membrane level, S-palmitoylation of FAK [318], a key EMT effector, facilitates its membrane localization. Meanwhile, N-myristoylation of p22, a Ca²⁺- binding protein, facilitates microtubule attachment, promoting intracellular trafficking essential for directional migration [319]. Intriguingly, the formins FMNL2 and FMNL3, which regulate spindle migration, rely on N-myristoylation to optimize cytoskeletal flexibility, reinforcing actin polymerization for cell shape adjustments [320, 321].

Innate immune responses

Growing evidence suggests that the innate and adaptive immune systems contribute to tumor progression when they are present in the TME [2]. Innate immune cells, including macrophages, dendritic cells (DCs), natural killer (NK) cells, and B cells, play crucial roles in tumor immunity within the TME, the innate immune system serves as the first line of defense against pathogens through pattern recognition receptors (PRRs). PRRs detect pathogen-associated molecular patterns (PAMPs) and damage-associated molecular patterns (DAMPs). This recognition triggers the production of proinflammatory cytokines and immunomodulatory molecules. PRRs are expressed primarily in antigen-presenting cells, such as DCs and macrophages [322,323,324].

Tumor-associated macrophages (TAMs)

Macrophages are a major component of the innate immune system in the TME and can be polarized into pro-inflammatory (M1-like) or immunosuppressive (M2-like) phenotypes, influencing tumor progression. Lipidation regulating macrophage function by modulating TLR signaling, cytokine production, and phagocytosis.

Lipopolysaccharide (LPS) from gram-negative bacteria activates the Toll-like receptor 4 (TLR4) signaling pathway, which involves four adaptor molecules: myeloid differentiation primary response 88 (MyD88), MyD88 adapter-like (MAL), Toll/IL- 1 receptor domain-containing adapter inducing IFN-β (Trif), and Trif-related adapter molecule (TRAM) [325]. The N-myristoylation of TRAM is critical for the activation of TLR4 signaling, which, upon LPS stimulation, facilitates TRAM’s translocation to the PM and its phosphorylation by protein kinase C ε (PKCε). This leads to the recruitment of Trif, activating nuclear factor kappa B (NF-κB) and interferon regulatory factor 3 (IRF3), promoting the pro-inflammatory response [326, 327] (Fig. 4). As a myristate-binding proteinheme oxygenase (HO- 2) negatively regulates the TRAM-dependent LPS‒TLR4 pathway, modulating the intensity of the immune response [328]. Beyond TLR signaling, N-myristoylation of ARF1 inhibits the cyclic GMP-AMP synthase (cGAS)-STING pathway, reducing the production of type I interferons (IFN-I) and dampening antitumor immunity [223]. Meanwhile, S-palmitoylation of STING aggregated at the Golgi apparatus, enhance IFN-I response, which, depending on strength and duration, may either promote tumor elimination or support immune evasion [329, 330] (Fig. 5). In contrast, S-palmitoylation of cGAS inhibits its binding to DNA, negatively regulating antitumor immune responses [331]. The S-palmitoylation of upstream proteins [332, 333] affects IFN-I responses, and downstream proteins depend on S-palmitoylation and influence antitumor immunity. Upon activation, IFN I activates the JAK-STAT pathway through the IFN-alpha receptor subunit (IFNAR1) to amplify immune responses [334]. Like a checkpoint, S-palmitoylation of IFNAR1 accelerates its degradation to prevent sustained activation and activates STAT2, indirectly affecting the transcription of STAT1 and IFN-α [335]. S-palmitoylation also affects the membrane localization of GPCR family members(e.g., CCR5 [336] and Rac1 [337]), influencing macrophage responses to inflammatory chemical signals. S-prenylation is also essential for maintaining macrophages' normal anti-inflammatory functions. By promoting the interaction between KRAS and PI3 K, the PI3 K pathway is activated, which regulates TLR signaling and moderates excessive inflammatory responses [338]. For the long isoform of the ZAP antiviral protein, farnesylation is essential for its targeting endolysosomes, thereby increasing its antiviral activity [339] (Fig. 6).

Phagocytosis, another central function of macrophages, is also influenced by lipidation. S-palmitoylation enhances Fcγ receptor-mediated phagocytosis in macrophages, with ZDHHC6-dependent lipidation increasing the efficiency of receptor clustering [340]. Similarly, CD36, a scavenger receptor involved in lipid uptake, relies on S-palmitoylation for its localization, further supporting macrophage activation [341]. Moreover, bacterial-sensing receptors such as nucleotide-binding oligomerization domain (NOD1/2) receptors require S-palmitoylation for activation, ensuring an effective immune response to microbial and damage-associated signals [342, 343]. Geranylgeranylation is critical for cytoskeletal dynamics, insufficient GGPP impairs the ability of monocytes and macrophages to respond to chemokines [344].

DCs

The impact of lipidation extends beyond macrophages to DCs, which act as crucial antigen-presenting cells, bridging the innate and adaptive immune response. N-myristoylation regulates endocytosis in DCs, as demonstrated by the HIV- 1 protein Nef, which prevents DC internalization of antigens by upregulating DC-SIGN expression, disrupting antigen processing and immune synapse formation [345, 346]. Similarly, S-palmitoylation enhances antigen uptake and presentation by facilitating the internalization of palmitoylated peptides, which are then processed and loaded onto major histocompatibility complex(MHC) I molecules, accelerating the immune activation process, making S-palmitoylation a powerful tool for antigen enrichment [347]. Moreover, S-palmitoylation modulates the production of IFN-α by activating interferon regulatory factor 7, ensuring proper DC inflammatory responses [348]. Blocking TLR2 S-palmitoylation inhibits cell surface expression and diminishes inflammatory responses to microbial ligands [349]. Structural organization within DCs is also lipidation-dependent; cDC1 s utilize their unique cross-presentation ability to degrade engulfed extracellular proteins and present them through MHC I molecules, activating CD8⁺ T cells and initiating antitumor immune responses. Geranylgeranylated Rac1 is necessary for actin filament reorganization, these proteins are then recognized by CLEC9 A (a C-type lectin receptor) on cDC1s, further activating CD8⁺ T cells and strengthening antitumor immunity (Fig. 6) [275, 350].

NK cells

The function of NK cells, which provide rapid cytotoxic responses to tumor cells, is also regulated by lipidation. S-palmitoylation enhances the clustering of NK receptors such as NKG2D and 2B4 with their ligands (MICA/MICB) within cholesterol-rich microdomains, optimizing NK cell activation and tumor cell recognition [351, 352] (Fig. 7). Additionally, S-palmitoylation ensures syntaxin 11 localization at immune synapses, maintaining its interaction with Munc18 - 2 and promoting lysosome exocytosis, thereby enhancing NK cell-mediated cytotoxicity [353].

Adaptive immune responses

B cells

B cells, essential component of the immune response, rely on lipidation for the proper function of B-cell receptor (BCR) signaling. N-myristoylation and S-palmitoylation of Lyn and human germinal center-associated protein (HGAL) stabilizes these proteins within lipid raft microdomains, enhancing their interactions with downstream effectors [354,355,356] (Fig. 4). This dual lipidation-dependent clustering facilitates the phosphorylation of immunoreceptor tyrosine-based activation motifs (ITAMs) by Lyn, activating spleen tyrosine kinase (SYK), Bruton’s tyrosine kinase (BTK), and phospholipase Cγ (PLCγ), which amplify downstream signals via the NF-κB, PI3 K/AKT, MAPK/extracellular signal-related kinase (ERK), and nuclear factor of activated T cells (NFATs) pathways [354,355,356]. In parallel, S-palmitoylation enhances BCR signaling by promoting the aggregation of BCR complexes with CD19/CD21/CD81 within lipid raft domains, prolonging signal transduction and increasing B-cell activation [357]. However, during B-cell apoptosis, caspase- 3/7 mediated cleavage of Lyn removes its N-myristoylation and S-palmitoylation sites, leading to its delocalization from the membrane, thereby preventing c-Myc regulation and suppressing B-cell survival [358]. Pan-NMT inhibitors block the N-myristoylation of Lyn, HGAL, and ARF1, accelerating their degradation and effectively inhibiting BCR signal transduction [354].

T cells

T-cell development and function are critically dependent on lipidation. In the TME, disruptions in S-palmitoylation impair key signaling proteins, such as the linker for activation of T cells (LAT), leading to T-cell anergy and immune dysfunction. Restoring LAT S-palmitoylation, alongside Lck expression, has been suggested as a strategy to reverse T-cell exhaustion and enhance antitumor responses [359]. Similarly, N-myristoylation is essential for maintaining T-cell homeostasis, as its deficiency results in aberrant T-cell development due to impaired Lck function, leading to a reduced number of CD4⁺ and CD8⁺ single-positive T cells in lymphoid organs [360]. Beyond development, NMT1 deficiency disrupts metabolic signaling by suppressing the AMPK pathway while hyperactivating TORC1, skewing T-cell differentiation toward proinflammatory TH1 and TH17 subsets and driving excessive IFN-γ and IL- 17 production, which may exacerbate autoimmune responses [361]. Interestingly, γδ T cells, which serve as immune sentinels against tumors, exhibit a unique independence from thymic N-myristoylation, though its inhibition disrupts the calcineurin–NFAT pathway and paradoxically enhances IFN-γ secretion [362, 363]. The depletion of GGPP induces stress responses in monocytes, triggering caspase- 1 activation and IL- 18 release, which in turn drives γδ T-cell activation [364].

T-cell activation is precisely regulated by S-palmitoylation and N-myristoylation, which work in concert to stabilize key immune receptors and fine-tune TCR signal transduction (Fig. 3). S-palmitoylation prevents the degradation of CD80, maintaining its stability at the immune synapse and promoting costimulatory interactions essential for full T-cell activation [365]. Additionally, it governs the localization of CD4, CD8, and LAT in cholesterol-rich membrane microdomains, ensuring that TCR signaling components are properly assembled [366,367,368]. Following TCR engagement, S-palmitoylation drives the phase separation of Grb2, Gads, SOS, and Slp76, leading to the formation of distinct signaling clusters that amplify downstream activation cascades [369]. Crucially, the activation of the TCR complex depends on the lipidation of Src family kinases, particularly Lck and Fyn, which require both N-myristoylation and S-palmitoylation for proper localization and function [370, 371]. Lck, a central mediator of TCR signaling, relies on these dual lipid modifications to interact with ZAP70 and PLC-γ1, initiating phosphorylation cascades that reinforce T-cell activation [372]. Likewise, Fyn undergoes dual lipidation to enhance ITAM phosphorylation on CD3ζ, facilitating the recruitment of Lck and amplifying TCR signaling [373, 374]. Once recruited to the TCR complex, ZAP70 phosphorylates LAT and SLP76, a process that is critically dependent on S-palmitoylation to stabilize these proteins within signaling microdomains [375, 376]. The transport of LAT-containing vesicles to the immune synapse is further regulated by ZDHHC18, highlighting a lipidation-dependent mechanism that sustains TCR responses [377]. While LAT has long been associated with lipid raft domains, emerging evidence suggests that its stability and plasma membrane anchoring are primarily dictated by S-palmitoylation, serving as a sorting signal for its localization [378, 379]. This process is linked to CD59, which appears to assist in the transfer of palmitoyl groups to LAT [380, 381], and prevent LAT degradation, thus prolonging TCR signaling duration and ensuring sustained immune activation [382, 383].

The role of S-prenylation in TCR signaling is further exemplified by CRACR2 A, a Rab GTPase that requires S-prenylation for its proper localization near the Golgi apparatus, ensuring effective control of calcium influx, JNK signaling, and vesicle trafficking [384]. These processes collectively support immune synapse formation and enhance the precision of T-cell activation (Fig. 6).

Immune checkpoints

PD- 1 is expressed on activated T cells, while PD-L1 is found on tumor and immune cells, where their interaction modulates immune responses and contributes to immune evasion [385] (Fig. 5). In tumor cells, S-palmitoylation of PD-L1 acts as a critical molecular switch regulating its stability and function. DHHC3-mediated S-palmitoylation of PD-L1 prevents its monoubiquitination, blocking its degradation through the endosomal sorting complex required for transport pathway and lysosomal degradation [21]. This enhances PD-L1 interaction with PD- 1, suppressing T-cell antitumor responses and promoting immune evasion. Inhibition of S-palmitoylation promotes PD-L1 degradation and restores T-cell activity, suggesting a potential therapeutic approach to counter PD-L1-mediated immune suppression [21].

Although DHHC3 mediates PD-L1 S-palmitoylation in colorectal cancer (CRC) cells, ZDHHC9 is identified as the primary S-palmitoylation enzyme for PD-L1 in other cancers, lung adenocarcinoma [268], PDAC [386] and BrCa [387], ZDHHC9-mediated S-palmitoylation stabilizes PD-L1, promoting immune evasion. A lack of ZDHHC9 shifts the TME from an immunosuppressive (“cold”) state to a proinflammatory (“hot”) state, primarily through CD8 + T-cell-dependent mechanisms, increasing the sensitivity of tumors to anti-PD-L1 immunotherapies [386]. Increased S-palmitoylation of PD-L1 is associated with chemotherapy resistance in tumor cells [388].

Similarly, S-palmitoylation of PD- 1, mediated by ZDHHC9, prevents its lysosomal degradation and enhances its stability, contributing to immune evasion. Palmitoylated PD- 1 also activates the mTOR pathway, promoting tumor cell proliferation. In three-dimensional culture systems, targeting PD- 1 S-palmitoylation has shown significant antitumor effects [138]. By inhibiting PD- 1 S-palmitoylation, the accumulation of PD- 1 on the tumor cell surface is reduced, thereby enhancing T-cell-mediated tumor immune responses. Targeting PD- 1 and PD-L1 S-palmitoylation, combined with immune checkpoint inhibitors or CAR-T cell therapies, may help overcome the existing clinical drug resistance. T-cell immunoglobulin and mucin domain-containing protein 3 (TIM- 3) is an emerging immune checkpoint, and its high expression exacerbates dysfunction and exhaustion in CD8 + T cells and NK cells while blocking TIM- 3 enhances their antitumor function [389]. The S-palmitoylation of TIM- 3 is regulated by DHHC9, which inhibits the interaction between TIM- 3 and HRD1, preventing its ubiquitination and degradation, thus stabilizing the TIM- 3 protein [390].

Thus, ZDHHC9 plays a crucial role in the tumor immune microenvironment, particularly in immune checkpoint regulation. Therefore, inhibiting ZDHHC9 emerges as a promising anticancer strategy. Given the emerging resistance to current immune checkpoint inhibitors in clinical settings, developing inhibitors targeting S-palmitoylation, such as ZDHHC9 selective inhibitors, is particularly important, and preclinical studies have demonstrated their efficacy [138]. Furthermore, the combination of immune checkpoint inhibitors with CAR-T therapy has shown significant therapeutic efficacy in recent studies [391], in a study of 11 mesothelioma patients who underwent preconditioning with cyclophosphamide, followed by a single dose of mesothelin-targeted CAR-T cells and at least three doses of anti-PD- 1 therapy, two patients achieved a 72% response rate with complete metabolic responses [392], further validating the potential of this combinatorial approach. Notably, research has also revealed surprising effects from the combination of ZDHHC9 inhibition and CAR-T therapy, such as the attenuation of CAR-T cell exhaustion upon DHHC9 knockdown and the acceleration of TIM- 3 degradation through a TIM- 3 S-palmitoylation peptide inhibitor, which enhances the antitumor immune response mediated by CAR-T cells and NK cells [390].

Cancer-associated fibroblasts

Cancer-associated fibroblasts(CAFs) in the TME are metabolically reprogrammed to support tumor growth by supplying fatty acids and modulating lactate metabolism. Lipidation, particularly S-palmitoylation, regulates multiple aspects of fatty acid uptake, storage, and metabolic signaling in these stromal cells, influencing tumor progression and immune evasion. regulate many aspects of fatty acid uptake and storage in these stromal cells.

A key regulator of fatty acid metabolism in CAFs is CD36, a membrane receptor that mediates the uptake of long-chain fatty acids. CD36 undergoes S-palmitoylation on intracellular cysteine residues, which is essential for its efficient localization to the PM and lipid rafts. In these membrane microdomains, palmitoylated CD36 forms complexes with Src-family kinases like Fyn and Lyn, facilitating fatty acid internalization and metabolic signaling [393]. Disrupting CD36 S-palmitoylation causes its retention in intracellular compartments, reduces fatty acid uptake, and attenuates downstream metabolic signaling, such as AMPK inhibition. TAMs [394] and CAFs [395] both express CD36, suggesting that heightened CD36 S-palmitoylation in these stromal cells promotes fatty acid influx for metabolic adaptation. Additionally, CD36 mediates oxidized LDL uptake-dependent macrophage migration inhibitory factor (MIF) expression, which recruits CD33 + myeloid-derived suppressor cells (MDSCs) and contributes to immune suppression and immunotherapy resistance [395].

In the Reverse Warburg Effect, glycolytic CAFs serve as the primary lactate producers in the TME, exporting lactate via monocarboxylate transporter 4 (MCT4), which is then taken up by tumor cells through MCT1. This lactate shuttle fuels tumor metabolism and enhances invasion by activating the TGF-β1/p38 MAPK/MMP2/9 signaling pathway [396].

Beyond metabolism, MAGUK p55 subfamily member (Mpp7) is a membrane-palmitoylated protein associated with CAFs that influences cancer progression by regulating the Rap1 signaling pathway [397]. Moreover, Mpp6 mediates the protumorigenic activity of serum amyloid A3 and stimulates the growth of pancreatic cancer cells [398]. In addition, S-palmitoylation inhibits receptor tyrosine kinase-mediated MAPK pathways by anchoring Spry proteins, thereby inhibiting cell proliferation and differentiation induced by FGF and VEGF and limiting the excessive activation of fibroblasts [399]. S-palmitoylation also regulates the membrane localization of the Rho family GTPase Wrch- 1, which promotes anchorage-dependent growth transformation in fibroblasts [400]. It induces large-scale membrane restructuring and massive endocytosis under mitochondrial stress conditions, thus maintaining membrane integrity and function [401].

Cancer-associated endothelial cells

eNOS is a central mediator of pro-angiogenic signaling, as it produces nitric oxide (NO) to induce vasodilation and support vessel sprouting [402,403,404,405,406]. eNOS is one of the best-characterized dually lipid-modified proteins: it undergoes co-translational N-myristoylation and post-translational S-palmitoylation on its N-terminus, these modifications target eNOS to the Golgi and then to plasma membrane caveolae in endothelial cells (ECs) [407, 408]. S-palmitoylation of eNOS at two N-terminal cysteines (Cys15 and Cys26) is mediated by DHHC21 and is reversible [409]. Proper localization of eNOS in caveolae (which are rich in the palmitoylated protein caveolin- 1 [410]) is required for its optimal activation by calcium-CaM [411] and Akt signaling [412]. Mutations that prevent eNOS N-myristoylation [413] or S-palmitoylation [414] cause eNOS to mislocalize to the cytosol and drastically reduce NO production. This leads to impaired angiogenic responses since NO is a pro-survival and pro-migratory signal for ECs. Studies have shown that agonist such as bradykinin stimulation of ECs triggers dynamic depalmitoylation of eNOS, causing it to temporarily leave caveolae and generate NO in other membrane compartments [415]. This cycle of S-palmitoylation/depalmitoylation is thought to fine-tune NO release during blood vessel remodeling. The importance of eNOS coupling further underscores by metabolic S-palmitoylation: ECs rely on endothelial cell-intrinsic FASN to supply palmitate for eNOS modification. When FASN is knocked out in endothelium, eNOS S-palmitoylation is lost, and reparative angiogenesis is impaired [416]. Thus, maintaining eNOS lipidation is a crucial aspect of tumor angiogenesis.

Another component is VEGF signaling. The primary VEGF receptor on endothelium, VEGFR1, which can act as a decoy or modulator of VEGF, exists in a membrane-bound form (mVEGFR1) that is palmitoylated, which serves as a “binary switch” for its stability: palmitoylated VEGFR1 is retained on the PM and recycled, whereas depalmitoylation targets VEGFR1 for internalization and lysosomal degradation [417]. Rab27a, a prenylated small GTPase, was identified as a critical upstream regulator of this process which help Traffick the S-palmitoylation enzymes [417]. Inhibition of FTase reduces the expression of hypoxia-inducible factor 1-alpha (HIF1α), prevents the binding of HIF1α to heat shock protein 90, and decreases the production of VEGF, thereby inhibiting angiogenic activities [418]. Additionally, other endothelial proteins required for vessel integrity are palmitoylated, for example, platelet endothelial cell adhesion molecule 1 is S-palmitoylated, which promotes its cell-surface localization in ECs [419]. Aberrant S-palmitoylation of such adhesion molecules could lead to the dysfunctional, leaky vessels characteristic of tumors. Beyond adhesion molecules, S-palmitoylation also influences endothelial motility and angiogenesis. The S-palmitoylation of RhoJ, regulated in a GLUL-dependent manner, enhances EC motility and promotes angiogenesis [420]. Similarly, the PI3 K–AKT–mTOR pathway, which drives angiogenesis and cell proliferation in clear cell renal cell carcinoma, can be activated through the S-palmitoylation of AGK [175]. In contrast, S-palmitoylation can also exert anti-angiogenic effects. For example, activation of the Src/p53 signaling pathway via S-palmitoylation of KAI1 inhibits angiogenesis [421].

For new blood vessels to form and infiltrate a growing tumor, endothelial cells must proliferate, migrate, and invade the surrounding matrix. These processes are driven by cytoskeletal remodeling and cell polarity pathways that heavily rely on prenylated small GTPases like Rho, Rac, and Cdc42 [422]. Active Rac1 drives lamellipodia formation for cell migration [423], while RhoA and RhoB influences stress fibers [424] and cell contraction [425]. In tumor angiogenesis, excessive Rho/Rac signaling can lead to abnormal vessel structure and hyper-permeability, aiding tumor cell intravasation. Pharmacological studies with statins which block the MVA pathway and thus S-prenylation illustrate the role of these GTPases: simvastatin was found to interfere with angiogenic sprouting by preventing Rho geranylgeranylation, thereby inhibiting Rho-mediated actin changes, effects reversible by adding back geranylgeranyl pyrophosphate [422].

Integrate role of lipidation in the TME

Protein lipidation does not act in isolation; rather, it forms an intricate signaling network that unifies immune evasion [21], metabolic reprogramming [426], and angiogenesis [416] within the TME. Tumor cells exploit lipidation to modulate immune signaling and evade immune surveillance [427]; simultaneously, immune and stromal cells leverage lipidation to respond dynamically to tumor-derived metabolic cues, such as lactate [428, 429], maintaining a complex and delicately balanced TME (Fig. 7).

Lipidation critically regulates TCR and BCR signaling by ensuring the proper membrane localization and activation of Src-family kinases (SFKs), such as Lck [372] and Fyn [373, 374] in T cells, and Lyn [354,355,356] in B cells. This lipidation-driven membrane localization amplifies downstream signaling cascades involving ZAP70 [372], LAT [376] in TCR signaling, and SYK, BTK, and PLCγ2 [354,355,356] in BCR signaling, thereby activating pathways like NF-κB and NFAT [354,355,356]. Indeed, BCR-mediated antigen recognition is not only the initiating signal for intracellular signaling in B cells but also regulates MHC II antigen presentation through human leucocyte antigen HLA-DM/HLA-DO, thereby enhancing T cell activation [430]. After TCR activation, cytotoxic T lymphocytes form an immune synapse with tumor cells, leading to the release of perforin, which creates membrane pores, allowing granzymes to enter and trigger tumor cell apoptosis [431]. However, tumor cells exploit lipidation to promote immune evasion. ZDHHC9-mediated S-palmitoylation stabilizes PD- 1 [138], which recruits Src homology region 2 domain-containing phosphatases. This inhibits TCR and CD28 signaling, weakening T cell activation, antigen recognition, and cytotoxic function, thereby sustaining T cell exhaustion and allowing tumors to escape immune surveillance [432].

In addition, tumor cells drive the differentiation of regulatory B cells (Bregs), which secrete IL- 10 and TGF-β while suppressing IFN-γ, TNF, and MCP- 1 production, this shifts the TME toward immunosuppression, inhibiting CD8⁺ and CD4⁺ T cell activation and impairing TCR-mediated cytotoxic responses [433, 434]. Bregs promote Treg differentiation [435] which the immune suppression effect was further reinforced. Within Treg, FTase promotes the maintenance of effector Tregs, GGTase I is a rheostat for TCR-dependent transcriptional programming and Rac-mediated signaling, establishing eTreg cell differentiation and immune tolerance [436].

Conversely, lipidation can enhance antitumor responses by optimizing innate immune receptor clustering. NK cells rely on S-palmitoylation to enhance receptor-ligand interactions within lipid rafts, such as NKG2D-MICA/MICB and 2B4-CD48, facilitating tumor cell recognition and cytotoxicity [351, 352]. S-palmitoylation of MyD88 by ZDHHC6 is required to recruit TLR complexes, driving NF-κB activation and proinflammatory cytokine release [437]. N-myristoylation of TRAM directs its localization to PM microdomains, which phosphorylated by PKCε facilitates TRIF-mediated IFN-I signaling in response to TLR4 activation [326, 327]. However, S-palmitoylation of IFNGR1 serving as a lysosomal sorting signal [427], which is driven by receptor degradation, reduces interferon responsiveness, diminishing antigen presentation and enabling immune escape and suppressing MHC class I upregulation, allowing tumor cells to escape CD8⁺ T-cell-mediated cytotoxicity [427].

The TME is characterized by metabolic reprograming, where lipidation and lactate metabolism converge to regulate cellular fitness. Palmitoylation of lactate LDHA enhances glycolytic flux, leading to lactate accumulation, a hallmark of the Warburg effect [16, 428]. This lactate surplus acidifies the TME, impairing T cell activation by inhibiting signaling pathways such as NFAT, p38, and JNK [438], while simultaneously stabilizing HIF1α [439], a master regulator of tumor metabolic adaptation. Beyond acidification, lactate acts as an intercellular metabolite, fueling metabolic coupling among tumor, stromal, and immune cells. Well-oxygenated tumor cells and CAFs internalize lactate via MCT1, utilizing it as an energy source while freeing glucose for hypoxic tumor regions—a process termed the "reverse Warburg effect" [429]. Lipidation contributes to this metabolic reprogramming by modulating key metabolic enzymes, supporting tumor proliferation and angiogenesis.

Simultaneously, lactate modifies immune signaling via G-protein coupled receptor 81 (GPR81), leading to cAMP suppression and PD-L1 upregulation, further reinforcing immune evasion [440]. DCs express GPR81, whose lactate-mediated activation suppresses the cell-surface expression of MHC-II, thereby compromising the ability of DCs to present tumor-cell antigens to T cells [441]. In cancer associate macrophages, this can shift them to an M2-like state that supports tumor growth [442]. This cross-regulation ensures a coordinated immune-metabolic landscape.

Crucially, lactate’s effects extend beyond immune modulation. Elevated lactate stabilizes HIF1α under normoxic conditions, promoting VEGF production [443]. In ECs, lactate uptake via endothelial MCT1 activates NF-κB/IL- 8 signaling, driving migration, vessel formation, and angiogenesis [444]. Simultaneously, lactate-driven metabolic stress triggers CAF activation through mitochondrial dysfunction, mitophagy, and subsequent secretion of growth factors such as hepatocyte growth factor(HGF) [445], promoting therapy resistance.

In essence, protein lipidation serves as a molecular currency traded between tumor and its microenvironment – it anchors immunomodulatory synapses, drives metabolic coupling, and tunes angiogenic signals in a unified, TME-wide manner.

Therapeutic implications and drug development

Developing lipidation-targeting drugs presents unique challenges in clinical trial design, largely due to the broad impact of lipidation on multiple signaling pathways. A key hurdle is patient selection, as these agents influence fundamental protein modifications rather than single oncogenic drivers. Early-phase clinical trials must adopt biomarker-driven strategies, selecting patients based on lipidation profiles (e.g., high NMT1 expression, low NMT2 expression) to ensure therapeutic efficacy [446].

Another significant challenge is compensatory signaling, which can undermine treatment efficacy. Lessons from early FTI trials highlight the importance of pathway redundancy, as blocking farnesylation alone was ineffective unless targeting HRAS-mutant tumors [447]. Regulators now emphasize the need for trial designs that account for alternative modifications, such as KRAS switching to geranylgeranylation when farnesylation is inhibited [49]. To demonstrate target engagement, patient tumor samples should be analyzed in real-time.

Toxicity remains a key concern, given that lipidation is a fundamental cellular process. However, due to the innovative nature of lipidation-targeting therapies, several first-in-class agents, such as NMT inhibitor PCLX- 001 [448] and TEAD S-palmitoylation inhibitor ISM6331 [142], received fast track or orphan drug designations to expedite development. Early-phase clinical trials should begin with conservative dosing, escalating gradually while closely monitoring adverse events and organ function (e.g., liver enzymes, bone marrow counts) on a weekly basis.

AI is increasingly leveraged in biomarker analysis, drug candidate selection, and safety monitoring, improving target specificity while reducing off-target toxicity. Recent AI-driven models have refined the development of lipidation inhibitors, such as ISM6331, by predicting binding affinities and optimizing molecular selectivity [142, 449]. However, the ethical challenges associated with AI-driven drug development—such as data privacy, cybersecurity risks, and regulatory gaps—necessitate clear guidelines and informed consent protocols [450, 451].

For lipidation-targeting therapies to be widely accessible, clinical trials must eliminate socioeconomic biases by providing free or subsidized companion diagnostics for key biomarkers (e.g., ZDHHC9/20 status). Diverse patient enrollment and long-term follow-up are essential to evaluate therapeutic outcomes and identify late-onset toxicities. Ultimately, the integration of real-time safety monitoring, validated biomarkers, and adaptive dose modifications will be crucial to balancing therapeutic efficacy with patient safety.