Lactylation-driven stabilization of the S100A11/ANXA2 complex promotes pancreatic cancer metastasis

Abstract

Pancreatic ductal adenocarcinoma (PDAC) is characterized by a pronounced Warburg effect and high lactate levels. While lysine lactylation (Kla) is an emerging post-translational modification, its role in regulating the S100A11/ANXA2 complex-a critical driver of membrane repair and metastasis-remains unexplored. In this study, we identified that S100A11 (K3, K55) and ANXA2 (K49) are significantly lactylated in PDAC tissues. Lactate-induced lactylation of S100A11 at K3 and K55 inhibits FBXW11-mediated ubiquitination and proteasomal degradation, thereby stabilizing the protein. This stabilized S100A11 subsequently shields ANXA2 from TRIM21-mediated degradation. Reciprocally, lactylation of ANXA2 at K49 enhances its capacity to recruit S100A11 to the plasma membrane. Functional assays demonstrated that this lactylation-driven interdependent regulatory axis promotes PDAC cell migration, invasion, and metastasis both in vitro and in vivo. Clinical analysis of TCGA data revealed that low expression of both S100A11 and ANXA2 predicts significantly improved progression-free and overall survival in PDAC patients. In summary, our findings establish a novel link between metabolic reprogramming and the post-translational regulation of the membrane repair machinery. Lactylation of the S100A11/ANXA2 axis is a key driver of PDAC progression and serves as a promising prognostic biomarker and potential therapeutic target.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s10565-026-10225-9.

Introduction

Pancreaticductal adenocarcinoma (PDAC) remains one of the most lethal malignancies, characterized by aggressive local invasion, early metastasis, and a dismal five-year survival rate of less than 12% (Rozengurt and Eibl 2026). A defining feature of the PDAC microenvironment is the pronounced Warburg effect, in which cancer cells preferentially utilize aerobic glycolysis, generating copious amounts of lactate even in the presence of sufficient oxygen (Liu et al. 2023, 2025). Recent studies have unveiled lactate’s role as a signaling molecule and substrate for a novel post-translational modification-lysine lactylation (Kla) (Ren et al. 2025; Zhang et al. 2026). The acetyltransferase p300 serves as the predominant canonical writer to catalyze lysine lactylation across diverse non-histone proteins in tumor cells; meanwhile, several deacylases including SIRT2, SIRT3, HDAC2, HDAC3 and HDAC8 have been validated as core erasers to reverse lactylation modification dynamically (Wang et al. 2026; Zhang et al. 2025). These well-characterized lactylation-modifying enzymes universally dominate the addition and removal of lactate moieties on most cytoplasmic and membrane-associated functional proteins in glycolysis-enriched tumor microenvironments (Ren et al. 2025; Wang et al. 2026; Zhang et al. 2025).

The S100 protein family comprises a group of small, calcium-binding proteins that participate in a wide spectrum of intracellular and extracellular functions, including regulation of protein phosphorylation, cytoskeletal dynamics, calcium homeostasis, and cell motility (Zeng et al. 2022; Zhuang et al. 2021). Among them, S100A11 has emerged as a significant tumor marker in PDAC, with elevated expression correlating with worse prognosis, enhanced pentose phosphate pathway activity, and increased malignant behavior, including aggressive invasion and metastasis (Zeng et al. 2022; Zhuang et al. 2021). Notably, S100A11 can form a heterotetrameric complex with Annexin A2 (ANXA2) in a calcium-dependent manner, a partnership that is critical for plasma membrane repair-a process upon which invasive cancer cells heavily depend to survive the mechanical stresses encountered during migration and metastasis (Ashraf and Gerke 2022; Koerdt and Gerke 2017; Lauritzen et al. 2015). Biochemical, structural, and functional analyses have revealed that this interaction relies on the N-terminal residues of ANXA2 (Rintala-Dempsey et al. 2006), which insert into a hydrophobic binding groove formed by the C-terminal extension (residues 93–105) of the S100A11 dimer (Ashraf and Gerke 2022). Multiple mutation assays, including deletion of the C-terminal annexin-binding region of S100A11 or disruption of its calcium-binding capacity, have been rigorously validated to abrogate this interaction, confirming the specific interacting interface between these two proteins (Ashraf and Gerke 2022; Liu et al. 2015; Rintala-Dempsey et al. 2006).

ANXA2 is a multifunctional phospholipid-binding protein that is frequently overexpressed in PDAC and has been associated with tumor progression, immune evasion, and poor clinical outcomes (Christofidis et al. 2024; Liu et al. 2025). Functionally, ANXA2 promotes membrane repair by facilitating the recruitment of lipid vesicles to sites of membrane injury, and its inhibition impairs cancer cell dissemination in preclinical models (Gounou et al. 2023; Lauritzen et al. 2015). Crucially, the robust accumulation of ANXA2 required for these pro-metastatic functions relies on overcoming its basal degradation. The steady-state stability of ANXA2 is known to be strictly regulated by the ubiquitin–proteasome pathway, where the E3 ubiquitin ligase TRIM21 plays a central role in targeting it for degradation. The interplay between S100A11 and ANXA2 at the plasma membrane is therefore of particular interest, as it may represent a critical node linking metabolic reprogramming to the membrane dynamics that enable metastasis.

Despite the established roles of S100A11 and ANXA2 individually in PDAC, several key questions remain unanswered. Whether and how these proteins are regulated by lactylation, how such modifications influence their ubiquitin-mediated turnover, interaction, and subcellular localization, and whether lactylation-dependent regulation of the S100A11/ANXA2 complex contributes to metastatic progression are all unknown. Given that the lactate-rich tumor microenvironment of PDAC provides an abundant substrate for Kla modifications, we hypothesized that lactylation may serve as a critical post-translational mechanism governing the S100A11/ANXA2 axis in pancreatic cancer.

In this study, we identify specific lactylation sites on S100A11 (K3 and K55) and ANXA2 (K49) in pancreatic cancer tissues and cell lines. Lactylation of S100A11 inhibits the E3 ligase FBXW11-mediated ubiquitination and proteasomal degradation, thereby stabilizing S100A11, which in turn shields ANXA2 from TRIM21-mediated degradation. Concurrently, lactylation of ANXA2 at K49 enhances its capacity to recruit S100A11 to the plasma membrane. This lactylation-driven interdependent regulatory axis promotes pancreatic cancer cell migration, invasion, and metastasis both in vitro and in vivo. Furthermore, clinical analysis reveals that low expression of both ANXA2 and S100A11 predicts a highly favorable prognosis in PDAC patients. Our findings establish a previously unrecognized link between metabolic reprogramming and post-translational regulation of the membrane repair machinery in PDAC.

Materials and methods

Detailed experimental procedures, reagents, and primer sequences are provided in the Supplementary Methods.

Results

S100A11 and ANXA2 are lactylated in pancreatic cancer tissues compared to adjacent normal tissues

To systematically characterize the lactylation landscape of S100 family proteins in pancreatic cancer, we analyzed published lactylome MS data from pancreatic cancer patients. Multiple Kla sites were detected across two S100 family members, with S100A6 at K47, and S100A11 at K3 and K55 showing prominent lactylation (Fig. 1A). S100A11 can interact with ANXA2 and modulate Ca2 + signaling-related membrane repair (Ashraf and Gerke 2022; Jaiswal et al. 2014; Lauritzen et al. 2015). Therefore, we also examined ANXA2 lactylation in these tissues. Notably, ANXA2 also exhibited significantly increased Kla levels at K49 in pancreatic cancer samples (Fig. 1A).

Fig. 1.

S100A11 and ANXA2 are lactylated and show increased membrane localization and interaction in response to lactate in pancreatic cancer. AHeatmap showing normalized lysine lactylation (Kla) intensities of S100 family proteins and ANXA2 in pancreatic cancer (PAT1–PT10) and adjacent normal tissues (PN1–PN10) based on published MS data (Duan et al. 2024). Red arrows indicate significantly different Kla sites between normal and tumor tissues, including S100A6 (K47), S100A11 (K3, K55), and ANXA2 (K49).B Immunofluorescence images of PANC-1 and MIA PaCa-2 cells treated with 0 or 20 mM L-lactate, stained for S100A11 (red), ANXA2 (green), and DAPI (blue). Scale bars, 10 μm. C, D Co-immunoprecipitation of S100A11 and ANXA2 in PANC-1 (C) and MIA PaCa-2 (D) cells treated with increasing concentrations of lactate (0, 5, 20 mM). Input and IP samples were analyzed by Western blotting for S100A11 and ANXA2. Oversaturated whole cell lysates were incubated with a fixed amount of antibody to ensure equal pull-down of the bait protein. E, H Western blot analysis of S100A11 and ANXA2 in membrane and whole cell lysates (WCL) from PANC-1 (E) and MIA PaCa-2 (H) cells treated with lactate. ATP1A1 and β-actin serve as membrane and loading controls, respectively. F, G, I, J Quantification of relative membrane (F, I) and whole cell (G, J) protein levels of S100A11 and ANXA2 in PANC-1 (F, G) and MIA PaCa-2 (I, J) cells. Data are mean ± SD (n = 3); p < 0.05, p < 0.01, p < 0.001 by one-way ANOVA

Lactate promotes membrane localization and interaction of S100A11 and ANXA2

The subcellular distribution of S100 proteins is related to their roles in the activation of different intracellular and extracellular cascades (Raeisi et al. 2025; Xue et al. 2026). Therefore, we examined whether the subcellular distribution of the S100A11/ANXA2 complex is affected by lactate. Immunofluorescence analysis revealed increased membrane localization and colocalization of S100A11 and ANXA2 in PANC-1 and MIA PaCa-2 cells upon lactate treatment (Fig.1B). Co-immunoprecipitation experiments confirmed enhanced interaction between S100A11 and ANXA2 in response to increasing lactate concentrations in both cell lines (Fig. 1C, D). Western blot analysis of subcellular fractions demonstrated that lactate treatment increased the membrane and whole cell levels of S100A11 and ANXA2 in a dose-dependent manner (Fig. 1E, H). Quantification confirmed significant upregulation of both proteins in membrane and whole cell fractions following lactate exposure (Fig. 1F, G, I, J).

S100A11 undergoes lactylation at K3 and K55, while ANXA2 is lactylated at K49 in pancreatic cancer cells

Co-IP assays showed that both S100A11 and ANXA2 exhibited dose-dependent increases in Kla modification upon lactate treatment (Fig. 2A, B). Importantly, quantitative RT-PCR analysis showed that lactate exposure did not significantly alter the mRNA levels of S100A11 or ANXA2, indicating that the observed changes in lactylation were not due to transcriptional regulation (Fig. 2C, D).

Fig. 2.

S100A11 undergoes lactylation at K3 and K55, while ANXA2 is lactylated at K49 in pancreatic cancer cells. A, B Co-IP analysis of S100A11 and ANXA2 lactylation in PANC-1 (A) and MIA PaCa-2 (B) cells treated with increasing concentrations of L-lactate (0, 5, 20 mM). Kla modification was detected using anti-Kla antibody. C, D Quantitative RT-PCR analysis of S100A11 and ANXA2 mRNA levels in PANC-1 (C) and MIA PaCa-2 (D) cells following lactate treatment for 24 h. E, F Co-IP analysis of lactylation of flag-tagged S100A11 WT, K3R, K55R, and K3R/K55R mutants in PANC-1 (E) and MIA PaCa-2 (F) cells treated with 20 mM lactate. G, H Co-IP analysis of lactylation of flag-tagged ANXA2-WT and K49R mutant in PANC-1 (G) and MIA PaCa-2 (H) cells treated with 20 mM lactate. I, J Co-IP analysis of the interaction between Flag-tagged S100A11 WT and K3R/K55R mutant with FBXW11 in PANC-1 (I) and MIA PaCa-2 (J) cells treated with 0 or 20 mM lactate. K, L Co-IP analysis of the interaction between Flag-tagged ANXA2-WT and K49R mutant with TRIM21 in PANC-1 (K) and MIA PaCa-2 (L) cells treated with 0 or 20 mM lactate

To pinpoint the specific lysine residues responsible for lactylation, we generated flag-tagged S100A11 wild-type (WT), K3R, K55R, and K3R/K55R mutants and expressed them in the two cell lines. Co-IP analysis revealed that lactylation of S100A11 was markedly reduced in both K3R and K55R single mutants, and almost completely abolished in the K3R/K55R double mutant, confirming that K3 and K55 are the major lactylation sites in S100A11 (Fig. 2E, F). Similarly, mutation of K49 to arginine (K49R) in ANXA2 abrogated its lactylation in both cell lines, demonstrating that K49 is the primary lactylation site in ANXA2 (Fig. 2G, H).

S100A11 can be degraded via FBXW11-mediated ubiquitination and proteasomal degradation (Chen et al. 2025; Huang et al. 2025), while the S100A11 and ANXA2 interaction can reduce ANXA2 proteasomal degradation (Tu et al. 2019). Therefore, we examined whether lactylation at these sites affects the interaction of S100A11 and ANXA2 with their respective E3 ubiquitin ligases. Co-IP experiments showed that lactate treatment reduced the interaction between S100A11-WT and FBXW11. Importantly, the non-lactylatable S100A11-K3R/K55R mutant retained its basal ability to interact with FBXW11, and crucially, this interaction was no longer abrogated by lactate treatment (Fig.2I, J). To rule out the possibility that the mutations nonspecifically abolished the binding interface due to conformational changes, we performed 3D structure predictions using AlphaFold 3, which confirmed that the K3R/K55R substitutions did not alter the overall folding of S100A11 (Supplementary Fig. 1). Furthermore, immunofluorescence assays demonstrated that lactate treatment did not independently alter the endogenous expression or subcellular localization of FBXW11 (Supplementary Fig. 2). For ANXA2, the interaction with TRIM21 was examined in WT and the K49R mutant (Fig. 2K, L). Lactate treatment reduced the interaction between TRIM21 and ANXA2-WT, and a similar reduction was observed for the ANXA2-K49R mutant (Fig. 2K, L). Thus, although ANXA2 K49 is the major lactylation site, its modification does not appear to influence the TRIM21-ANXA2 interaction, suggesting that the lactylation-mediated regulation of E3 ligase binding is specific to S100A11.

Lactylation of S100A11 K3 and K55 inhibits FBXW11-mediated ubiquitination and proteasomal degradation

To investigate whether lactylation at K3 and K55 affects S100A11 protein stability, we performed cycloheximide (CHX) chase assays in PANC-1 and MIA PaCa-2 cells expressing Flag-tagged S100A11 wild-type (WT) or the K3R/K55R mutant, with or without lactate treatment. In both cell lines, S100A11-WT protein levels gradually decreased over time after CHX addition, indicating ongoing protein degradation (Fig. 3A, B). Lactate treatment significantly delayed the degradation of S100A11-WT, suggesting that lactylation enhances S100A11 stability. In contrast, the K3R/K55R mutant showed similar degradation rates regardless of lactate exposure, indicating that K3 and K55 are required for lactate-induced stabilization of S100A11 (Fig. 3C, D). Quantification confirmed that lactate markedly prolonged the half-life of S100A11-WT but not the K3R/K55R mutant in both PANC-1 and MIA PaCa-2 cells (Fig. 3E, F). In addition, S100A11 K3R/K55R mutant had prolonged half-life compared to S100A11-WT (Fig. 3E, F), suggesting that K3 and K55 are themselves targets of poly-ubiquitination, and that lactylation at these sites may directly compete with ubiquitin conjugation to protect S100A11 from degradation.

Fig. 3.

Lactylation of S100A11 at K3 and K55 inhibits FBXW11-mediated ubiquitination and proteasomal degradation. A–D Cycloheximide (CHX) chase assays in PANC-1 (A, C) and MIA PaCa-2 (B, D) cells expressing Flag-tagged S100A11 wild-type (WT) or K3R/K55R mutant, treated with or without 20 mM L-lactate. Cells were harvested at the indicated time points after CHX addition, and S100A11 protein levels were analyzed by Western blotting. E, F Quantification of relative S100A11 protein levels over time in PANC-1 (E) and MIA PaCa-2 (F) cells. G, H Ubiquitination assays in PANC-1 (G) and MIA PaCa-2 (H) cells expressing Flag-S100A11-WT alone or in combination with FBXW11, after treated with 20 mM L-lactate, and/or MG132 as indicated. Ubiquitinated S100A11 was detected by immunoblotting (IB) with anti-ubiquitin antibody following Flag immunoprecipitation. I, J Ubiquitination assays in PANC-1 (I) and MIA PaCa-2 (J) cells expressing Flag-S100A11-WT or -K3R/K55R mutant, treated with 20 mM L-lactate, and/or MG132 as indicated

Given that S100A11 is known to undergo FBXW11-mediated ubiquitination and proteasomal degradation (Chen et al. 2025), we next examined whether lactylation at K3 and K55 affects S100A11 ubiquitination. Co-immunoprecipitation and ubiquitination assays revealed that lactate treatment significantly reduced FBXW11-mediated polyubiquitination of S100A11-WT in both PANC-1 and MIA PaCa-2 cells (Fig.3G, H). In contrast, the K3R/K55R mutant was less ubiquitinated regardless of lactate treatment (Fig. 3I, J). Therefore, lactylation of S100A11 at K3 and K55 could inhibit FBXW11-mediated ubiquitination and subsequent proteasomal degradation, thereby stabilizing S100A11 protein in pancreatic cancer cells.

S100A11 alleviates TRIM21-mediated ANXA2 proteasomal degradation

It’s known that TRIM21 can promote ANXA2 proteasomal degradation (Liu et al. 2024). To investigate whether S100A11 regulates TRIM21-mediated ANXA2 degradation in pancreatic cancer cells, we first examined the interaction between ANXA2 and TRIM21 in the presence or absence of S100A11. Co-immunoprecipitation assays revealed that knockdown ofS100A11 significantly increased the binding of TRIM21 to ANXA2 in both PANC-1 and MIA PaCa-2 cells (Fig. 4A-C). This inverse relationship strongly suggests a competitive binding mechanism, wherein the presence of S100A11 restricts TRIM21 from accessing ANXA2. We next assessed the effect of S100A11 on ANXA2 protein stability. Western blot analysis showed that S100A11 knockdown markedly reduced ANXA2 protein levels, and this reduction was rescued by treatment with the proteasome inhibitor MG132, but not by the lysosome inhibitor BafA1, in both cell lines (Fig. 4D-G). These results indicate that S100A11 protects ANXA2 from proteasomal, but not lysosomal, degradation.

Fig. 4.

S100A11 attenuates TRIM21-mediated ubiquitination and proteasomal degradation of ANXA2 in pancreatic cancer cells. A, B Co-immunoprecipitation (Co-IP) analysis of the interaction between ANXA2 and TRIM21 in PANC-1 (A) and MIA PaCa-2 (B) cells transduced with control shRNA (shNC) or S100A11 shRNA (shS100A11). Input and IP samples were immunoblotted for TRIM21, ANXA2, S100A11, and β-actin. C Quantification of relative TRIM21 binding to ANXA2 in PANC-1 and MIA PaCa-2 cells (n = 3). D-G Western blot analysis of ANXA2 and S100A11 protein levels in PANC-1 (D) and MIA PaCa-2 (F) cells with or without S100A11 knockdown, treated with 20 mM lactate, 10 μM MG132, or 100 nM BafA1 as indicated. E, G Quantification of relative ANXA2 protein levels in PANC-1 (E) and MIA PaCa-2 (G) cells (n = 3). H–K Western blot analysis of ANXA2 and S100A11 protein levels in PANC-1 (H) and MIA PaCa-2 (J) cells with S100A11 and/or TRIM21 knockdown, treated with or without 20 mM lactate as indicated. I, K Quantification of relative ANXA2 protein levels in PANC-1 (I) and MIA PaCa-2 (K) cells (n = 3). L, M Ubiquitination assays in PANC-1 (L) and MIA PaCa-2 (M) cells with S100A11 and/or TRIM21 knockdown, treated with 20 mM lactate and 10 μM MG132 as indicated. Ubiquitinated ANXA2 was detected by immunoblotting with an anti-ubiquitin antibody following ANXA2 immunoprecipitation. *p < 0.05, **p < 0.01, ***p < 0.001, ns, not significant

Knockdown of TRIM21 successfully restored ANXA2 protein levels in S100A11-deficient cells under both basal conditions and upon lactate treatment. Lactate-induced ANXA2 upregulation was abolished by S100A11 knockdown, but this reduction was completely reversed when TRIM21 was simultaneously depleted (Fig. 4H-K), indicating that S100A11 stabilizes ANXA2 by antagonizing TRIM21-mediated degradation. Finally, we examined the effect of S100A11 and TRIM21 on ANXA2 ubiquitination. Ubiquitination assays demonstrated that lactate treatment reduced ANXA2 polyubiquitination. S100A11 knockdown enhanced ANXA2 polyubiquitination, while simultaneous knockdown of TRIM21 abolished this effect in both cell lines (Fig. 4L, M). Collectively, these findings indicate that S100A11 inhibits TRIM21-mediated ubiquitination and proteasomal degradation of ANXA2 in pancreatic cancer cells.

ANXA2 is required for S100A11 membrane localization and its lactylation enhances this effect

Having established that S100A11 stabilizes ANXA2 by antagonizing TRIM21-mediated degradation, we next investigated whether ANXA2 reciprocally regulates S100A11 subcellular localization. Co-immunoprecipitation analysis in MIA PaCa-2 cells revealed that S100A11 interacted with both Flag-tagged ANXA2 wild-type (WT) and the K49R mutant, indicating that lactylation at K49 does not affect the physical association between S100A11 and ANXA2 (Fig. 5A).

Fig. 5.

ANXA2 is required for S100A11 membrane localization and its lactylation enhances this effect. A Co-immunoprecipitation analysis of interaction between S100A11 and Flag-tagged ANXA2 wild-type (WT) or K49R mutant in MIA PaCa-2 cells. S100A11 was immunoprecipitated and the presence of ANXA2 was detected by immunoblotting with anti-Flag antibody. B Immunofluorescence analysis of S100A11 (red) and Flag-ANXA2 (green) in MIA PaCa-2 cells with stable ANXA2 knockdown (shANXA2) reconstituted with Flag-ANXA2-WT or K49R mutant, treated with or without 20 mM L-lactate for 24 h. Nuclei were stained with DAPI (blue). Scale bars, 10 μm. C, D Western blot analysis of S100A11 and ANXA2 in membrane fractions from PANC-1 (C) and MIA PaCa-2 (D) cells transduced with control shRNA (shNC) or ANXA2 shRNA (shANXA2), treated with or without 20 mM L-lactate for 24 h. E, F Western blot analysis of S100A11 and ANXA2 in membrane fractions, alongside ANXA2 in whole cell lysates (WCL) from PANC-1 (E) and MIA PaCa-2 (F) cells with stable ANXA2 knockdown (shANXA2) reconstituted with vector (Vec), Flag-ANXA2-WT, or K49R mutant, treated with or without 20 mM L-lactate for 24 h

To determine whether ANXA2 and its lactylation influence S100A11 membrane localization, we established stable ANXA2-knockdown PANC-1 and MIA PaCa-2 cells and reconstituted them with Flag-tagged ANXA2-WT or K49R mutant. Immunofluorescence analysis demonstrated that in cells reconstituted with ANXA2-WT, lactate treatment markedly enhanced S100A11 membrane localization and its colocalization with ANXA2 (Fig. 5B). In contrast, cells expressing the ANXA2-K49R mutant showed diminished S100A11 membrane enrichment even after lactate exposure, suggesting that ANXA2 lactylation at K49 is important for promoting S100A11 membrane targeting (Fig. 5B).

Subcellular fractionation experiments further confirmed these findings. Knockdown of ANXA2 substantially reduced S100A11 levels in the membrane fraction of both PANC-1 and MIA PaCa-2 cells, and lactate treatment failed to rescue S100A11 membrane localization in the absence of ANXA2 (Fig. 5C, D). Importantly, reconstitution with ANXA2-WT restored S100A11 membrane levels and re-established the lactate-induced enhancement of S100A11 membrane localization in both cell lines (Fig. 5E, F). Notably, western blot analysis of the membrane fractions also revealed that lactate treatment strongly promoted the membrane translocation of ANXA2-WT itself, whereas the ANXA2-K49R mutant showed impaired membrane accumulation (Fig. 5E, F). As ANXA2 is known as an intrinsically membrane-binding protein involved in membrane repair (Koerdt and Gerke 2017), and its interaction with S100A11 remains intact regardless of the K49 mutation (Fig.5A), these data suggest a co-translocation mechanism. Specifically, lactylation at K49 enhances ANXA2’s own membrane affinity, enabling it to act as a carrier that shuttles the associated S100A11 to the cell membrane.

Consistently, re-expression of the ANXA2-K49R mutant only partially restored S100A11 membrane association and did not support further enhancement upon lactate stimulation (Fig. 5E, F). Collectively, these data demonstrate that ANXA2 is required for S100A11 membrane localization in pancreatic cancer cells, and that lactylation of ANXA2 at K49 further potentiates this membrane-targeting effect by driving the entire complex to the membrane.

ANXA2 promotes pancreatic cancer cell migration, invasion, and metastasis in an S100A11- and lactylation-dependent manner

S100A11/ANXA2 complex contributes to plasma membrane repair, which is important for cancer cell survival under mechanical stress (Koerdt and Gerke 2017; Lauritzen et al. 2015). In addition, S100A11 and ANXA2 not only facilitate membrane repair but also promote tumor cell motility and metastatic potential (Tu et al. 2019). Therefore, we next investigated whether ANXA2 and its lactylation at K49 contribute to pancreatic cancer cell migration, invasion, and metastasis.

To address this, we performed gain- and loss-of-function experiments in PANC-1 and MIA PaCa-2 cells. ANXA2 and S100A11 knockdown (shANXA2 and shS100A11) significantly impaired both migration and invasion in vitro, as assessed by transwell assays (Fig. 6A-D). Restoration of wild-type ANXA2 (ANXA2-WT) rescued the migratory and invasive abilities of ANXA2-deficient cells, whereas re-expression of the lactylation-deficient mutant (ANXA2-K49R) only partially restored these phenotypes. Given that highly glycolytic pancreatic cancer cells constantly produce basal levels of endogenous lactate, these results indicate that ANXA2 K49 lactylation driven by endogenous lactate is required for full pro-metastatic function. Importantly, simultaneous knockdown of S100A11 (shS100A11) in ANXA2-WT reconstituted cells markedly attenuated the rescue effect, suggesting that S100A11 is necessary for ANXA2-mediated cell migration and invasion (Fig. 6A-D).

Fig. 6.

ANXA2 promotes pancreatic cancer cell migration, invasion, and metastasis in an S100A11- and lactylation-dependent manner. A, B Representative images of transwell migration and invasion assays in PANC-1 (A) and MIA PaCa-2 (B) cells with the indicated genetic manipulations. Migrated or invaded cells were stained and imaged (scale bar, 100 μm). C, D Quantification of relative cell migration and invasion percentages in PANC-1 (C) and MIA PaCa-2 (D) cells. E Schematic of the in vivo metastasis assay. MIA PaCa-2 cells with the indicated genetic modifications were injected into the tail vein of 6–8-week-old female BALB/c nude mice (n = 8 per group). After 4 weeks, lungs were harvested for histological analysis. F) Representative H&E-stained lung sections showing metastatic nodules (upper panels, low magnification; lower panels, high magnification of boxed areas). G Quantification of metastatic tumor burden in the lungs, expressed as the percentage of tumor area relative to the total lung area. **p < 0.01, *** p < 0.001, ns: not significant

To further evaluate the role of ANXA2 and S100A11 in metastasis in vivo, we utilized a tail vein injection model in BALB/c nude mice. MIA PaCa-2 cells with the indicated genetic manipulations were injected intravenously, and lung metastases were assessed after four weeks (Fig. 6E). Consistent with the in vitro findings, ANXA2 or S100A11 knockdown significantly reduced metastatic tumor burden in the lungs compared to the vector control; while re-expression of ANXA2-WT restored metastatic potential, the ANXA2-K49R mutant only partially restored it (Fig. 6F, G). This in vivo phenotype further underscores the functional importance of lactylation, as the physiological tumor microenvironment is inherently lactate-rich. Notably, S100A11 knockdown in the context of ANXA2-WT reconstitution abrogated the pro-metastatic effect of ANXA2, further supporting the requirement of S100A11 for ANXA2-driven metastasis. Collectively, these results demonstrate that ANXA2 promotes pancreatic cancer cell migration, invasion, and metastasis in a manner dependent on both its K49 lactylation and the presence of S100A11.

Low expression of both ANXA2 and S100A11 predicts favorable prognosis in pancreatic cancer patients

To corroborate our mechanistic findings at the protein level, we analyzed mass spectrometry-based proteomics data from the CPTAC-PADC cohort (Cao et al. 2021). Consistent with our in vitro data, the protein expression levels of both ANXA2 and S100A11 were significantly upregulated in pancreatic tumor tissues (n = 141) compared to normal tissues (n = 74) (Supplementary Fig. 3A; p < 0.001). Furthermore, a significant positive correlation was observed between the protein levels of S100A11 and ANXA2 in tumor samples (Supplementary Fig. 3B; Spearman’s rho = 0.40, p < 0.001), supporting their synergistic interaction in vivo.

To determine the clinical relevance of our mechanistic findings, we stratified TCGA-PAAD patients into four groups based on high or low expression (median cutoff) of ANXA2 and S100A11. Baseline clinicopathological characteristics of the patients stratified by the dual-biomarker signature (Low ANXA2/Low S100A11 vs. Others) are summarized in Table 1. Kaplan–Meier analysis revealed significant differences in both progression-free interval (PFI) and overall survival (OS) among the four groups (Fig. 7A, B; log-rank p = 0.012 and p = 0.009, respectively). Notably, patients with low expression of both ANXA2 and S100A11 exhibited significantly longer PFI and OS compared to all other patients combined (Fig. 7C, D; PFI: HR = 1.700, 95% CI: 1.154–2.505, p = 0.007; OS: HR = 1.775, 95% CI: 1.177–2.675, p = 0.006). These results indicate that low expression of both ANXA2 and S100A11 predicts a favorable prognosis in pancreatic cancer.

Table 1.

Comparison of the clinicopathological characteristics of PAAD patients by ANXA2 and S100A11 stratification in TCGA-PAAD

Characteristics	Others	Low ANXA2 and low S100A11	p value
n	109	69
Pathologic_stage, n (%)			1.000
I/II	103 (95.4%)	65 (95.6%)
III/IV	5 (4.6%)	3 (4.4%)
Age_at_initial_pathologic_diagnosis, median (IQR)	66 (57, 74)	65 (58, 72)	0.462
Gender, n (%)			0.123
Male	65 (59.6%)	33 (47.8%)
Female	44 (40.4%)	36 (52.2%)
Histologic grade, n (%)			0.017
G2	59 (54.1%)	36 (52.2%)
G3	36 (33%)	12 (17.4%)
GX	1 (0.9%)	1 (1.4%)
G1	13 (11.9%)	18 (26.1%)
G4	0 (0%)	2 (2.9%)
Anatomic_neoplasm_subdivision, n (%)			0.943
Body of Pancreas	8 (7.3%)	6 (8.7%)
Head of Pancreas	86 (78.9%)	53 (76.8%)
Other	6 (5.5%)	5 (7.2%)
Tail of Pancreas	9 (8.3%)	5 (7.2%)
Family_history_of_cancer, n (%)			0.634
YES	38 (55.9%)	26 (60.5%)
NO	30 (44.1%)	17 (39.5%)
History_of_diabetes, n (%)			0.049
NO	74 (79.6%)	35 (64.8%)
YES	19 (20.4%)	19 (35.2%)

IQR, interquartile range; GX, grade cannot be assessed. Significant p values (p < 0.05) are highlighted in bold. Statistical significance was evaluated using the Chi-square test (or Fisher’s exact test) for categorical variables and the Mann–Whitney U test for continuous variables

Fig. 7.

Low expression of both ANXA2 and S100A11 predicts favorable prognosis in pancreatic cancer patients. A, B Kaplan–Meier analysis of progression-free interval (PFI, A) and overall survival (OS, B) in TCGA-PAAD patients stratified into four groups by high or low expression of ANXA2 and S100A11. Log-rank test p values are shown. C, D Kaplan–Meier analysis comparing patients with low expression of both ANXA2 and S100A11 (red) to all other patients (blue). Hazard ratios (HR), 95% confidence intervals (CI), and log-rank p values are indicated

To further investigate whether this dual-biomarker signature could serve as an independent prognostic factor, we performed univariate and multivariate Cox regression analyses. In the univariate analysis, the low ANXA2/low S100A11 signature was a significant protective factor for both PFI (Table 2) and OS (Table 3). After adjusting for confounding factors including age, gender, pathologic stage, histologic grade, and anatomic subdivision in the multivariate analysis, the dual-low signature exhibited a strong trend toward being an independent predictor for favorable PFI (HR = 0.675, 95% CI: 0.433–1.050, p = 0.081) and OS (HR = 0.664, 95%CI: 0.416–1.061, p = 0.087). Although strict statistical significance (p < 0.05) was not reached in the multivariate model (likely due to sample size limitations within the TCGA-PAAD cohort), these results collectively indicate that the low expression of both ANXA2 and S100A11 might serve as a prognostic indicator for favorable clinical outcomes in pancreatic cancer.

Table 2.

Univariate and multivariate Cox regression analyses of progression-free interval (PFI) in the TCGA-PAAD cohort

Characteristics	Total (n)	Univariate analysis		Multivariate analysis
		Hazard ratio (95% CI)	p value	Hazard ratio (95% CI)	p value
ANXA2/S100A11 classification	178
Others	109	Reference		Reference
Low ANXA2 and low S100A11	69	0.572 (0.378—0.865)	0.008	0.675 (0.433—1.050)	0.081
Pathologic stage	176
I/II	168	Reference
III/IV	8	0.983 (0.399—2.423)	0.970
Age at initial pathologic diagnosis	178	1.013 (0.995—1.033)	0.165
Gender	178
Male	98	Reference
Female	80	1.027 (0.698—1.510)	0.894
Histologic grade	178
G2	95	Reference		Reference
G3	48	1.469 (0.954—2.264)	0.081	1.549 (0.982—2.443)	0.060
GX	2	1.874 (0.257—13.671)	0.536	1.631 (0.223—11.908)	0.630
G1	31	0.573 (0.314—1.043)	0.068	0.739 (0.398—1.372)	0.338
G4	2	1.333 (0.184—9.670)	0.776	2.092 (0.258—16.947)	0.489
Anatomic subdivision	178
Body of Pancreas	14	Reference		Reference
Head of Pancreas	139	2.471 (1.072—5.694)	0.034	2.310 (0.998—5.344)	0.051
Other	11	0.640 (0.160—2.562)	0.529	0.566 (0.140—2.290)	0.425
Tail of Pancreas	14	2.011 (0.715—5.657)	0.186	1.621 (0.557—4.716)	0.376
Family history of cancer	111
Yes	64	Reference
No	47	1.036 (0.624—1.719)	0.892
History of diabetes	147
No	109	Reference
Yes	38	0.779 (0.458—1.324)	0.355

CI, confidence interval. Variables with p < 0.1 in the univariate analysis or with established clinical significance were included in the multivariate Cox regression model. GX, grade cannot be assessed

Table 3.

Univariate and multivariate Cox regression analyses of overall survival (OS) in the TCGA-PAAD cohort

Characteristics	Total (n)	Univariate analysis		Multivariate analysis
		Hazard ratio (95% CI)	p value	Hazard ratio (95% CI)	p value
ANXA2/S100A11 classification	178
Others	109	Reference		Reference
Low ANXA2 and low S100A11	69	0.543 (0.348—0.846)	0.007	0.664 (0.416—1.061)	0.087
Pathologic stage	176
I/II	168	Reference
III/IV	8	0.736 (0.232—2.336)	0.603
Age at initial pathologic diagnosis	178	1.028 (1.007—1.049)	0.009	1.021 (1.001—1.043)	0.041
Gender	178
Male	98	Reference
Female	80	1.214 (0.808—1.825)	0.350
Histologic grade	178
G2	95	Reference		Reference
G3	48	1.316 (0.836—2.071)	0.235	1.355 (0.845—2.172)	0.207
GX	2	0.880 (0.121—6.392)	0.900	0.571 (0.078—4.191)	0.582
G1	31	0.502 (0.259—0.975)	0.042	0.648 (0.330—1.273)	0.208
G4	2	0.837 (0.115—6.073)	0.861	0.944 (0.125—7.114)	0.956
Anatomic subdivision	178
Body of Pancreas	14	Reference		Reference
Head of Pancreas	139	2.438 (0.983—6.043)	0.054	2.338 (0.940—5.818)	0.068
Other	11	0.254 (0.030—2.177)	0.211	0.218 (0.025—1.907)	0.169
Tail of Pancreas	14	1.948 (0.616—6.156)	0.256	1.758 (0.544—5.678)	0.346
Family history of cancer	111
Yes	64	Reference
No	47	0.885 (0.517—1.517)	0.658
History of diabetes	147
No	109	Reference
Yes	38	0.923 (0.531—1.607)	0.778

CI, confidence interval. Variables with p < 0.1 in the univariate analysis or with established clinical significance were included in the multivariate Cox regression model. GX, grade cannot be assessed

In summary, based on our in vitro and in vivo findings, we propose a comprehensive mechanistic model illustrating how the lactate-enriched tumor microenvironment and endogenous glycolysis drive pancreatic cancer progression via the S100A11–ANXA2 axis (Fig. 8). Abundant intracellular lactate, sourced from both endogenous production and extracellular uptake, initially induces the lactylation of S100A11 at K3 and K55. This modification directly blocks the binding of the E3 ligase FBXW11, thereby rescuing S100A11 from ubiquitination and proteasomal degradation. The stabilized S100A11 then binds to ANXA2, acting as a physical shield that competitively restricts TRIM21 from mediating ANXA2 degradation. Concurrently, lactate induces the lactylation of ANXA2 at K49. This lactylated ANXA2 acts as a robust carrier, shuttling the stabilized S100A11–ANXA2 complex to the inner leaflet of the plasma membrane. Ultimately, the membrane-anchored accumulation of this complex potentiates membrane repair capacity and fuels the migration, invasion, and metastasis of pancreatic cancer cells.

Fig. 8.

Schematic model of the lactate-driven S100A11–ANXA2 regulatory axis in pancreatic cancer. Metabolic input from both endogenous glycolysis and the lactate-rich tumor microenvironment orchestrates a hierarchical stabilization and translocation cascade of the S100A11–ANXA2 complex. (1–2) Intracellular lactate induces the lactylation (Kla) of S100A11 at K3 and K55, which prevents its recognition by the E3 ligase FBXW11, thereby abolishing S100A11 polyubiquitination and proteasomal degradation. (3) The accumulated S100A11 binds to ANXA2, physically shielding it from TRIM21-mediated ubiquitination, which in turn stabilizes the ANXA2 protein. (4) Concurrently, lactate lactylates ANXA2 at K49. The K49-lactylated ANXA2 serves as a carrier to translocate the intact S100A11–ANXA2 complex to the inner leaflet of the plasma membrane. (5) The membrane-anchored complex ultimately executes its pro-malignant functions, including promoting membrane repair, cell migration, invasion, and tumor metastasis

Discussion

This study revealed a previously unrecognized mechanism by which metabolic reprogramming directly fortifies the metastatic machinery of PDAC cells. We demonstrate that lactate-induced lactylation of the S100A11/ANXA2 complex dictates its protein stability, subcellular localization, and pro-metastatic functions, establishing a direct mechanistic bridge between altered tumor metabolism and enhanced cancer cell dissemination.

Our findings significantly expand the current understanding of the functional scope of protein lactylation. Initial studies on lactylation predominantly focused on histone modifications and their role in epigenetic regulation (Hu et al. 2024). However, the lactylome is increasingly recognized to encompass a diverse array of non-histone proteins (Ren et al. 2025; Zhang et al. 2025). By identifying specific lactylation sites on S100A11 (K3 and K55) and ANXA2 (K49), we provide compelling evidence that lactylation directly regulates the cytoplasmic and membrane-associated protein networks governing cell motility and survival. This highlights a dynamic, metabolism-driven regulatory layer where the abundant lactate in the PDAC microenvironment is actively co-opted to enhance the mechanical resilience and invasive capacity of cancer cells.

A central discovery of our study is the intricate, lactylation-driven stabilization cascade involving S100A11, ANXA2, and their respective E3 ubiquitin ligases. We found that lactylation at K3 and K55 of S100A11 directly antagonizes its recognition and subsequent polyubiquitination by the E3 ligase FBXW11. This competitive interplay between lactylation and ubiquitination represents a crucial mechanism for extending protein half-life in a metabolite-dependent manner. Furthermore, the stabilization of S100A11 has downstream consequences for its binding partner, ANXA2. Our data indicate that S100A11 shields ANXA2 from TRIM21-mediated proteasomal degradation. Given that S100A11 knockdown significantly enhanced TRIM21-ANXA2 binding, this shielding effect is likely mediated by competitive interaction. Structurally, S100A11 interacts with the N-terminus of ANXA2 to form a robust heterotetramer (Ashraf and Gerke 2022; Liu et al. 2015; Rintala-Dempsey et al. 2006). We propose that the formation of this bulky complex either directly occupies the overlapping TRIM21 recognition site on ANXA2 or introduces substantial steric hindrance that physically masks the targeted lysine residues from the E3 ligase. Thus, the initial lactylation event on S100A11 initiates a stabilizing domino effect, ensuring the robust presence of the S100A11/ANXA2 complex within the cell.

Beyond protein stability, our study unveils the spatial regulation of this complex by lactylation. While ANXA2 lactylation at K49 does not alter its interaction with TRIM21, it is indispensable for the efficient recruitment of S100A11 to the plasma membrane. The S100A11/ANXA2 heterotetramer is a well-established executor of plasma membrane repair (Koerdt and Gerke 2017; Lauritzen et al. 2015; Prislusky et al. 2024), a process critically required by invading cancer cells to survive the severe mechanical shear stresses encountered during extracellular matrix traversal and vascular transit (Lauritzen et al. 2015). By demonstrating that ANXA2 lactylation enhances S100A11 membrane localization, we propose that the lactate-rich microenvironment essentially primes PDAC cells for metastasis by pre-assembling and fortifying their membrane repair machinery at the cell periphery. Although the precise biophysical mechanism behind this recruitment remains to be fully elucidated, we hypothesize that the addition of a bulky lactyl group—which neutralizes the positive charge of lysine 49—induces a local conformational shift in ANXA2. This structural modification may alleviate auto-inhibitory constraints, thereby exposing its core phospholipid-binding domains (e.g., increasing affinity for phosphatidylserine) or stabilizing the optimal conformation required for S100A11 tetramerization at the lipid bilayer.

Because S100A11 and ANXA2 function as an obligate complex to execute these processes, their pro-metastatic roles are strictly interdependent. In this highly coordinated system, S100A11 serves as the crucial metabolism-responsive element whose lactylation prevents the degradation of the complex, while ANXA2 functions as the structural, membrane-anchoring executor that translates this metabolic stabilization into enhanced invasive and membrane repair capacities. Similarly, the pro-metastatic effect of S100A11 intrinsically depends on both its own lactylation at K3/K55 (to prevent rapid proteasomal degradation) and the physical presence of ANXA2 (to shuttle S100A11 to the plasma membrane). We strategically focused our in vivo rescue experiments on ANXA2 because it serves as the terminal membrane-anchoring executor of this interdependent functional unit.

The clinical relevance of this metabolic-protein axis is underscored by our prognostic analysis of PDAC patients. The combined low mRNA expression of both ANXA2 and S100A11 serves as a robust predictor of highly favorable progression-free and overall survival. Given that S100A11 and ANXA2are established oncogenes, investigating the clinical significance of their co-expression is of critical importance. This dual-biomarker signature not only reflects the underlying biology of the tumor but also offers potential translational value for patient stratification. Furthermore, our mechanistic insights suggest that disrupting the S100A11/ANXA2 axis could represent a viable therapeutic strategy. While directly targeting protein–protein interactions remains pharmacologically challenging, intervening in the upstream lactylation machinery offers a more tractable approach (Sun et al. 2025; Zhang et al. 2026). Accumulated evidence indicates that the non-histone lactylome is dynamically controlled by conserved enzymes, with p300 acting as a canonical writer and SIRT2/3 or HDACs serving as major erasers in glycolysis-enriched microenvironments (Ren et al. 2025; Wang et al. 2026; Zhang et al. 2025). Identifying and inhibiting the specific lactyltransferases responsible for modifying S100A11 and ANXA2 could thus destabilize this pro-metastatic complex.

Despite these insights, our study has certain limitations that warrant further investigation. First, as we demonstrated that lactate treatment does not alter the mRNA levels of S100A11 and ANXA2, our current prognostic analysis using TCGA mRNA data serves only as a baseline indicator. Because the pro-metastatic function is driven by post-translational modifications, investigating the direct relationship between the specifically lactylated forms of S100A11 and ANXA2 and clinical prognosis in PDAC cohorts is a top priority for our future studies. The development of site-specific lactylation antibodies will be essential to accurately assess the clinical significance of their co-expression at the protein level. Second, the specific enzymes catalyzing the addition and removal of lactyl groups on S100A11 and ANXA2 remain to be identified. Uncovering these writers and erasers will be critical for fully understanding the dynamic regulation of this pathway. Third, while we focused on the S100A11/ANXA2 complex, the global impact of the PDAC lactylome on other membrane dynamics and cytoskeletal regulatory proteins requires broader exploration. Fourth, while our immunofluorescence data clearly establish a functional link between ANXA2 lactylation and S100A11 membrane recruitment, the exact biophysical mechanism remains speculative. Future structural biology studies, such as cryo-EM or in vitro liposome-binding assays, are required to ascertain whether K49 lactylation directly enhances ANXA2’s affinity for membrane phospholipids or structurally facilitates the assembly of the functional heterotetramer.

In conclusion, our study uncovers a novel paradigm in which tumor metabolism directly dictates the stability and spatial organization of the S100A11/ANXA2 complex via protein lactylation. This interdependent regulatory axis not only shields these proteins from proteasomal degradation but also promotes their membrane localization, thereby equipping PDAC cells with the enhanced migratory and membrane repair capabilities necessary for successful metastasis. Targeting this lactate-driven post-translational modification network may provide a new therapeutic strategy for combating metastatic progression in pancreatic cancer.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

B.Z. and L.C. performed the experiments, analyzed the data, and drafted the initial manuscript. Y.D. and L.Y. conceived the study, supervised the research, and provided critical revisions to the manuscript. All authors reviewed and approved the final version of the manuscript.

Funding

Department funded.

Data availability

Data are available from the corresponding authors on reasonable request.

Declarations

Ethical approval

All animal experiments were approved by the Basic and Clinical Research Ethics Committee of Sichuan Provincial People’s Hospital (approval no. 2026336) and performed in accordance with relevant guidelines.

Conflict of interests

The authors declare no competing interests.