Mast cell-derived MIF Polarizes Th17 cells to drive lung adenocarcinoma progression.

Introduction
Lung adenocarcinoma (LUAD) represents a leading cause of cancer-related mortality worldwide [1, 2], with tumor progression being critically regulated by complex cellular interactions within the tumor microenvironment (TME) [3]. While immune checkpoint inhibitors have revolutionized cancer treatment, a significant proportion of LUAD patients exhibit limited response [4], highlighting the need to better understand alternative immune-modulatory pathways in the TME.
Our previous research has demonstrated that mast cells (MCs) can be activated by tumor-derived exosomes to promote LUAD metastasis through tryptase-mediated mechanisms [5]. And MC exosomes can promote lung adenocarcinoma cell proliferation by the role of KIT-stem cell factor signaling [6]. Notably, the role of MCs in cancer is context-dependent, with studies reporting both pro-tumor [7–9] and anti-tumor [10–12]effects in lung cancer models. This functional duality may depend on factors such as tumor stage, microenvironmental cues, and the specific mediators released. Building on this foundation, we now seek to explore the broader immunomodulatory capacities of MCs within the LUAD microenvironment. These tissue-resident immune cells are capable of releasing a wide spectrum of pre-formed and newly synthesized mediators that can profoundly influence the inflammatory landscape [13, 14]. However, beyond their established pro-metastatic and pro- proliferating functions, the specific mechanisms through which MCs contribute to LUAD pathobiology, particularly their interactions with adaptive immune cells, remain incompletely defined.
Macrophage migration inhibitory factor (MIF), a pleiotropic cytokine constitutively expressed and rapidly secreted by various cell types including MCs, plays crucial roles in inflammation, immunity, oncogenesis [15, 16], and has been shown to promote Th17 cell responses [17–19]. Elevated MIF levels correlate with poor prognosis in multiple cancers, suggesting its pro-tumorigenic potential [20–22]. Concurrently, Th17 cells, defined by their production of interleukin-17A (IL-17A), have emerged as significant contributors to tumor-promoting inflammation and progression in multiple malignancies [23, 24]. Elevated Th17 cell activity has been associated with poor prognosis in several cancers, such as pancreatic [25], hepatocellular [26], pulmonary [27, 28], prostatic [29]. Furthermore, IL-17A is implicated in angiogenesis, metastasis, and the generation of pro-tumorigenic microenvironments across diverse human cancers [30], including lung [31], gastric [32], and breast [33] tumours.
Despite these parallel observations, the potential crosstalk between MC-derived MIF and Th17 cell polarization within the spatially organized LUAD TME remains largely unexplored. Spatial heterogeneity within the TME is increasingly recognized as a critical determinant of immune cell function, yet its impact on specific cellular circuits such as MC-Th17 interactions is poorly understood [34–36]. Furthermore, whether MIF serves as a critical molecular mediator linking MC activity to Th17-driven tumor promotion constitutes a significant knowledge gap in LUAD biology.
Based on these considerations, we hypothesized that MC-derived MIF orchestrates Th17 cell polarization within the LUAD TME, thereby establishing a novel immune regulatory axis that drives tumor progression. This study aimed to systematically investigate the "MC-MIF-Th17" axis through integrated retrospective clinical analysis, digital pathology analysis of immunofluorescence staining, single-cell transcriptomics, and functional validation, with the ultimate goal of identifying new therapeutic targets for LUAD treatment.

Results

Serum IL-17 is significantly elevated in LUAD patients with malignant pleural effusion (MPE)
This study enrolled 43 LUAD patients with malignant pleural effusion (MPE) and 21 non-MPE LUAD controls. Demographic and clinical characteristics of the study population are summarized in Supplementary Table S1. Several significant intergroup differences emerged, particularly in smoking history and tumor clinical staging. Comparative analysis of plasma cytokine profiles between groups is presented in Table 1. No statistically significant differences were observed in plasma concentrations of IL-1β, IL-2, IL-4, IL-5, IL-10, IL-12, IFN-γ, or TNF-α. However, the MPE group demonstrated markedly elevated IL-17 levels compared to controls (p < 0.001), aligning with prior reports of IL-17/MPE associations in LUAD. Additionally, significantly higher concentrations of IFN-α (p = 0.001), IL-8 (p = 0.002), and IL-6 (p = 0.01) were detected in MPE patients. To further validate the specificity of IL-17 elevation in MPE, we performed stratified analyses controlling for disease stage and treatment status. These analyses confirmed that serum IL-17 levels remained significantly higher in MPE patients than in non-MPE patients, irrespective of whether patients had received immunotherapy (anti-PD-1/PD-L1 treatment) or not (both p < 0.01). Furthermore, when comparing only stage IV patients, IL-17 levels were still markedly elevated in the MPE subgroup compared to the non-MPE subgroup (p < 0.001). Detailed results of these stratified comparisons are presented in Supplementary Table S2.

Quantitative and spatial correlation between MC and Th17 cell infiltration in human LUAD
To investigate the relationship between MCs and Th17 cells in the tumor microenvironment, we performed multiplexed immunofluorescence analysis on 80 human LUAD specimens. Representative imaging revealed the presence and spatial distribution of CK7 + tumor cells, Tryptase + MCs, and CD3 + CD4 + IL-17A + Th17 cells within the tissue architecture (Fig. 1A). The specificity of staining was further confirmed by reviewing the original single-channel images (Fig. S1). Subsequent digital image analysis generated protein expression heatmaps and in situ positive cell maps, providing a quantitative visualization of the cellular landscape (Fig. 1B). Quantitative analysis first revealed a significant positive correlation between the infiltration of these two cell types. Using non-parametric Spearman’ rank correlation analysis (appropriate for cell count data), we found consistent correlations for both the absolute numbers (Spearman’sρ = 0.550, p = 1.22e-07; Fig. 1C) and the relative proportions (Spearman’sρ = 0.506, p = 1.69e-06; Fig. 1D) of MCs and Th17 cells (data included in Table S3). To further assess the potential for direct cellular interaction, we quantified the spatial proximity between MCs and Th17 cells. The mean distance between each MC and its nearest Th17 neighbor within a 20 μm radius was significantly shorter in LUAD tumors (median: 8.3 μm, IQR: 11.6 μm) than in normal lung tissues (median: 10.9 μm, IQR: 13.1 μm). Notably, 63.7% of MC-Th17 cell pairs in tumors were found within 0–10 μm, a distance compatible with direct membrane contact or short-range paracrine signaling (data included in Table S4). The strong quantitative correlation and the close spatial proximity provide compelling foundational evidence for a potential functional interaction between mast cells and Th17 cells in the LUAD microenvironment.

MCs and Th17 cells are enriched and exhibit a distance-dependent co-localization in the peritumoral region
We next characterized the spatial distribution of MCs and Th17 cells relative to the tumor margin. Quantitative analysis revealed a significant, distance-dependent gradient, with the highest absolute density of both cell types located in the immediate peritumoral region (0–20 μm), progressively decreasing in the 20–40 μm and 40–60 μm zones (Fig. 2C, E). Intriguingly, while the absolute cell numbers were highest in the 0–20 μm zone, the relative proportion (percentage) of both MCs and Th17 cells peaked in the 20–40 μm zone (Fig. 2D, F), suggesting that the immediate tumor periphery, while most densely infiltrated, is also highly populated by other cell types. To robustly assess their co-infiltration pattern, we analyzed the correlation between MC and Th17 cell abundance across these zones using Spearman’s rank correlation. This analysis was based on tumor tissue samples from 80 patients, encompassing a total of 16,340 quantified cells (12,332 MCs and 4,008 Th17 cells) across all fields of view within the three peritumoral zones. The analysis revealed that the strength of this positive correlation exhibited a spatially graded decay: it was strongest in the 0–20 μm zone, moderate in the 20–40 μm zone, and weakest (though still significant) in the 40–60 μm zone (Fig. 2G-L). This stepwise decrease in correlation strength with increasing distance from the tumor margin reinforces the notion of a spatially organized, distance-dependent interaction between MCs and Th17 cells within the peritumoral niche. The complete spatial infiltration dataset is provided in Table S5.

Accumulation of MCs and Th17 cells is linked to disease progression and poor outcome
We next assessed the clinical relevance of MC and Th17 cell infiltration in LUAD (the complete dataset is provided in Table S3). Kaplan–Meier survival analysis revealed that patients with high levels of either MCs or Th17 cells, measured by both absolute abundance (Fig. 3A, B) and relative percentage (Fig. S2.A, B), experienced significantly worse overall survival. Furthermore, the infiltration density of both cell types was significantly higher in patients with advanced-pathological-stage (stages III and IV) disease compared to those with early-stage (stages I and II) disease, as well as in those with higher T categories (T3-4 vs. T1-2), positive nodal status (N2-3 vs. N0-1), and distant metastasis (M1 vs. M0). (Fig. 3C-F). To evaluate whether the observed prognostic association was independent of disease stage, we performed a stage-stratified survival analysis. Although statistical power was limited within each subgroup, the consistent trend—wherein higher infiltration levels of either cell type were associated with a tendency toward poorer survival—was maintained across most TNM stage subgroups (as shown in Fig. S2C-H). This suggests that the prognostic value of these cells extends beyond merely reflecting tumor stage. Together, these data establish that the accumulation of MCs and Th17 cells is a robust marker of tumor aggressiveness and poor prognosis.

Single-Cell RNA sequencing reveals a potential MC-MIF-Th17 cell crosstalk axis in the peritumoral niche
We analyzed the single-cell transcriptomic dataset of LUAD and peritumoral tissues. Following quality control, cell subsets were annotated based on established marker genes (Fig. 4A), with results visualized via the t-SNE algorithm. Through t-SNE-based dimensionality reduction clustering and classical marker identification, 12 major cell types were identified (Fig. 4B). Differential expression analysis was performed between cell subsets. To directly highlight the core cellular interaction axis of our study, a volcano plot specifically displaying the differentially expressed genes (DEGs) between MCs and Th17 cells is presented in Fig. 4C, with the top 10 significantly upregulated and downregulated DEGs annotated. The full multigroup comparison encompassing all cell subsets is provided in the Figure S3. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis was conducted on DEGs in MCs, comparing LUAD tumor tissue to adjacent normal tissue. Pathways with a significance level of P < 0.05, determined using the clusterProfiler R package, were considered significantly enriched (Fig. 4D). Notably, the DEGs in MCs showed significant enrichment in pathways related to Th17 cell differentiation signaling. Intercellular communication between MCs and other cell types in LUAD tumors versus adjacent tissues was analyzed using the CellPhoneDB database via the CellChat package. Bubble plots depict all interactions from MCs to other cell populations, including ligand-receptor pairs or signaling pathways. Compared to LUAD tissues, peritumoral tissues exhibited significantly upregulated interactions mediated by the MIF-(CD74 + CXCR4) ligand-receptor pair between MCs and Th17 cells (Fig. 4E, F). To assess the clinical relevance of MIF in systemic circulation, we prospectively measured serum MIF levels in an independent cohort of LUAD patients. Using ELISA, we found that serum MIF concentrations were significantly elevated in patients with distant metastasis (32.65 ± 8.93 ng/mL, n = 4) compared to those without metastasis (17.63 ± 6.10 ng/mL, n = 6) (unpaired t-test, p = 0.019). This positive correlation between circulating MIF levels and disease progression aligns with and extends our tissue-based findings, underscoring the translational potential of the MC-MIF-Th17 axis. These data are presented in Supplementary Figure S4.

LLC-conditioned medium (LCM) stimulates MIF secretion from MCs
Bone marrow-derived cells successfully developed into BMMCs, following a 4-week induction period in IL-3-containing culture conditions. Microscopic examination revealed abundant violet-hued cytoplasmic granules in these BMMCs. Flow cytometric validation targeting the characteristic MC surface markers CD117 and high-affinity IgE receptor (FcεR1) demonstrated differentiation efficiency exceeding 97.02% (Fig. 5A, B). The detailed gating strategy, established based on unstained controls to ensure rigorous threshold setting, is provided in Fig. S3A, B. And Microscopic examination revealed abundant violet-hued cytoplasmic granules in these BMMCs, characteristic heterochromatic particles can be seen by toluidine blue staining (Fig. 5C). To quantitatively delineate the effect of LLC-Conditioned Medium (LCM) stimulation on MIF, we performed qPCR and ELISA. At the transcriptional level, LCM stimulation led to an approximately fivefold increase in MIF mRNA compared to the unstimulated control. Pre-treatment with the stabilizer CS attenuated, but did not abolish, this upregulation (remaining > twofold) (Fig. 5D). At the protein level, ELISA confirmed a significant increase in MIF concentration in the supernatant of LCM-stimulated BMMCs (Fig. 5E). Western blot analysis of cell lysates and concentrated supernatants corroborated elevated MIF protein expression upon LCM stimulation alone, compared to controls and LCM + CS groups (Fig. 5F). Parallel analysis confirmed that the LCM stimulus itself contained no detectable MIF. These data demonstrate that LCM stimulation not only triggers the release of pre-formed MIF but, more importantly, actively promotes de novo MIF synthesis at the transcriptional level, leading to a potent amplification of MIF output from MCs.

MIF serves as an essential mediator for MC-driven Th17 polarization in the LUAD microenvironment
While MIF has been implicated in Th17 cell differentiation across various contexts, the specific role of MC-derived MIF in the LUAD tumor microenvironment remained unclear. To delineate this, we established a modified Th17-polarizing system and assessed the resulting Th17 cell differentiation by flow cytometry and qPCR. First, the addition of 0.5 µg/ml recombinant MIF to the differentiation medium significantly increased the proportion of Th17 cells among CD4⁺ T cells, an effect abolished by the MIF antagonist ISO-1 (200 µM) (Fig. 6A-D). To assess the contribution of MC-derived MIF, we then used conditioned media from LCM-stimulated BMMCs. Compared to the control medium, supernatant from LCM-stimulated BMMCs [SN(LCM)] robustly promoted Th17 differentiation (Fig. 6A-D). This effect was significantly attenuated, but not completely abolished, when the MIF inhibitor ISO-1 was added directly to the SN(LCM) culture [SN(LCM)/ISO-1] (Fig. 6E-H). Importantly, supernatant from BMMCs pre-treated with the membrane stabilizer CS) before LCM stimulation [SN(LCM/CS)] showed the weakest Th17-polarizing capacity, similar to baseline levels (Fig. 6 E–H). These results demonstrate that MIF is a key and necessary effector molecule in MC-mediated Th17 polarization, as its specific inhibition significantly impairs this process. However, the incomplete reversal by ISO-1 indicates that activated MCs also secrete additional pro-Th17 factors. Thus, MC-derived MIF works in concert with other mediators to drive Th17 cell polarization within the tumor microenvironment.

IL-17A promotes the proliferation, migration, and invasion of lung cancer cells
We next asked whether IL-17A produced by MIF-polarized Th17 cells could enhance tumor aggressiveness. Accordingly, LLC cells were treated with conditioned media from differentially polarized Th17 populations, and their proliferative, migratory, and invasive capacities were examined using colony formation, CCK-8, wound healing, and Transwell assays. Compared to the control, both exogenous MIF and conditioned medium from LCM-stimulated MCs significantly promoted Th17 differentiation and IL-17A secretion, enhancing malignant phenotypes in Lewis lung carcinoma (LLC) cells. These effects were inhibited by the MIF antagonist ISO-1 or the MC stabilizer CS (Fig. 7A-B, E, G-H, K-M). However, as CS globally inhibits MC secretion, the attenuated effect observed in the SN[Th17 + SN(LCM/CS)] group could not be attributed solely to the lack of MIF. To definitively establish the specific role of MC-derived MIF, we directly compared the functional output of Th17 cells polarized under three critical conditions designed to isolate the MIF effect: with MIF present (SN[Th17 + SN(LCM)]), with MIF specifically inhibited (SN[Th17 + SN(LCM)/ISO-1]), and with total MC secretion inhibited (SN[Th17 + SN(CS/LCM)]). The medium from Th17 cells polarized with sufficient MIF most potently promoted LLC colony formation, proliferation, migration, and invasion. Specific inhibition of MIF (ISO-1) during polarization significantly attenuated this pro-tumorigenic capacity, though residual activity remained. Th17 cells induced by CS-pretreated MCs produced medium with the weakest tumor-promoting effects (Fig. 7C-D, F, I-J, N-P). These results demonstrate that MC-derived MIF is a critical and necessary upstream driver for Th17 cells to acquire full pro-tumorigenic function. The residual activity upon specific MIF inhibition confirms that MCs secrete additional polarizing factors, but the absence of MIF alone substantially impairs the tumor-promoting activity of Th17 cells, establishing MIF as an indispensable core mediator in this pathway. Importantly, the pro-tumorigenic function of this MC-MIF-Th17 axis was successfully recapitulated in an independent murine lung adenocarcinoma cell line, CMT167. Conditioned media from MIF-polarized Th17 cells similarly enhanced the proliferation, migration, and invasion of CMT167 cells, effects that were likewise attenuated upon MIF inhibition (Figure S5). This consistency across two distinct cell lines underscores the generalizability of the identified mechanism.

Targeting the MC-MIF axis suppresses Th17 polarization and impedes lung cancer progression in vivo
Based on our preliminary experimental findings, we conducted in vivo studies to validate these observations, with the experimental design for animal investigations illustrated in Fig. 8A. Notably, both the MIF antagonist ISO-1 and the MC membrane stabilizer CS demonstrated significant inhibitory effects on Th17 polarization and lung cancer progression through blockade of MC-derived MIF. This was evidenced by reduced tumor fluorescence (Fig. 8B, C), as well as decreased terminal tumor volume and weight (Fig. 8E, F), without affecting overall mouse body weight (Fig. 8D). Mechanistically, both treatments markedly reduced Th17 cell infiltration within tumors, as shown by flow cytometry (Fig. 8G, H) using a gating strategy rigorously defined by fluorescence-minus-one (FMO) controls (Figue S4C, D) and by downregulation of IL-17A and RORγt mRNA (Fig. 8I, J). Immunohistochemistry confirmed a significant decrease in IL-17A⁺ cells in treated tumors (Fig. 8K, L), providing direct tissue-level evidence that targeting the MC-MIF axis inhibits Th17 polarization in the tumor microenvironment. Comparative analysis revealed that ISO-1, which directly antagonizes MIF, exhibited superior anti-tumor efficacy and greater suppression of Th17 cell differentiation compared to CS. This differential effectiveness suggests that MIF contributing to Th17 polarization and tumor progression may originate from cellular sources beyond MCs, potentially involving alternative MIF-secreting cell populations in the tumor microenvironment.

Discussion
This study, through the integration of clinical cohort analysis, digital pathology, single-cell transcriptomics, and functional experiments, systematically elucidates a novel mechanism by which MCs drive LUAD progression via MIF-mediated polarization of T helper 17 (Th17) cells. Our findings not only reveal the central role of the “MC-MIF-Th17” axis within the LUAD TME but also identify a promising therapeutic target for advanced LUAD (Fig. 9).
First, our clinical observations provided a solid starting point for this mechanistic pathway. Retrospective analysis revealed significantly elevated serum levels of interleukin-17A (IL-17A) in LUAD patients with malignant pleural effusion (MPE), suggesting a potentially important role for Th17 cells and their effector cytokine in LUAD malignant progression. It is noteworthy that our decision to focus subsequent mechanistic studies on IL-17A, among the multiple elevated cytokines identified in our retrospective analysis, was guided by the pursuit of a specific and biologically coherent hypothesis. Although IFN-α, IL-8, and IL-6 were all significantly elevated, their respective biological contexts pointed towards immunological pathways distinct from the MC-Th17 cell axis we sought to investigate. IFN-α is a hallmark of type I interferon responses, typically associated with antiviral and potential anti-tumor immunity, a direction contrary to our focus on pro-tumorigenic mechanisms [37]. IL-8 primarily orchestrates neutrophil-driven inflammation, representing a different leukocyte lineage [38]. Although IL-6 can contribute to Th17 cell differentiation, it is a pleiotropic upstream mediator with widespread effects, lacking the specificity required to directly implicate the Th17 effector phase [39]. In contrast, IL-17A serves as the definitive effector cytokine of Th17 cells [40]. Its elevation provided the most direct serological evidence linking the clinical observation to the testable hypothesis centered on MC-driven Th17 cell polarization. Subsequent histological studies confirmed the rationale for this choice. mIHC and digital pathology analysis demonstrated a significant positive correlation between MC infiltration density and Th17 cell abundance in LUAD tissues, and both were independent predictors of adverse patient prognosis, strongly suggesting a potential functional synergy between these two cell populations within the tumor locale.
Our in-depth analysis of the TME spatial architecture represents a key innovation of this study. We found that MCs and Th17 cells were not randomly distributed but were co-enriched in the peritumoral region, exhibiting a distinct distance-dependent co-localization pattern radiating from the tumor margin. This finding carries significant functional implications: the peritumoral area, acting as the “frontline” between the tumor and host tissue, is often the most active site for cytokine gradient formation, immune cell recruitment, and activation. The co-aggregation of MCs and Th17 cells in this niche creates an optimal microenvironment for their close cellular communication. Single-cell transcriptomic analysis provided molecular confirmation, clearly showing enhanced communication strength between MCs and Th17 cells in the peritumoral niche, with the MIF and its receptor pathways (e.g., CD74/CXCR family) identified as the key ligand-receptor pair mediating this interaction. This successfully links the clinically relevant cellular phenotypes with an underlying spatial regulatory mechanism.
Mechanistically, we established a functional axis initiated by tumor cells, relayed via MCs and MIF, and ultimately executed through Th17 cells. Critically, our in vitro polarization assays combining flow cytometry and functional readouts demonstrated that MC-derived MIF serves as the principal driver for Th17 differentiation. Although residual Th17 activity upon specific MIF inhibition by ISO-1—exceeding the effect of global MC secretion blockade by CS —indicates the involvement of other MC-derived auxiliary factors, the pronounced impairment of Th17 polarization upon MIF loss definitively establishes MIF as the non-redundant core mediator from MCs in this pathway. The polarized Th17 cells, via releasing their signature cytokine IL-17A, directly acted on tumor cells, enhancing their proliferative, migratory, and invasive capacities, thereby completing the final output from the immune microenvironment to malignant tumor behavior. It is worth noting that whereas prior studies, including our own, have established MC-derived mechanisms that directly enhance tumor cell behaviors—such as proliferation via exosomal KIT-stem cell factor signaling [7] and metastasis via tryptase release [6]—the present work unveils a distinct immunomodulatory pathway. Here, MCs promote LUAD progression indirectly by polarizing Th17 cells through MIF, thereby collectively outlining a multi-dimensional and multi-layered pro-tumor functional network orchestrated by MCs in LUAD.
The translational value of this study lies in our in vivo validation of the therapeutic potential of targeting the “MC-MIF-Th17” axis. In orthotopic lung cancer mouse models, intervention using either ISO-1 or cromolyn sodium significantly suppressed intratumoral Th17 cell polarization and effectively impeded tumor progression. This suggests two attractive clinical strategies: ISO-1 represents a precise strategy directly targeting the MIF factor [41]; whereas cromolyn sodium, an already approved drug for asthma treatment, offers the potential for rapid clinical translation via drug repurposing [42], aiming to reduce the release of MIF and other pro-tumor mediators at their source by stabilizing MC membranes.
Certainly, this study has limitations. First, regarding the experimental models, the BMMCs used herein represent a standardized in vitro model of the connective tissue-type MC. While this model is invaluable for mechanistic discovery due to its high yield and manipulability, we acknowledge that it may not fully recapitulate the phenotypic and functional heterogeneity of tissue-resident MCs within the human lung adenocarcinoma microenvironment. Therefore, extrapolating our findings to the in vivo setting warrants caution, and future validation using primary human MCs or in situ models would be highly informative. Our retrospective clinical cohort size warrants further expansion to strengthen the prognostic value of this axis. The specific upstream signaling pathways triggering MIF secretion from MCs, and the precise downstream mechanisms, such as epigenetic and metabolic reprogramming, by which MIF regulates Th17 cell differentiation, are important directions for future investigation. Furthermore, the role of this axis in other LUAD complications (e.g., cancer pain, cachexia) and its potential synergy with existing immune checkpoint inhibitor therapies represent highly valuable future research avenues.
In conclusion, this study systematically reveals a pro-tumor immune circuit driven by the “MC-MIF-Th17” axis within the LUAD TME. The elucidation of this circuit deepens our understanding of the immunomodulatory functions of MCs, provides novel biomarkers (co-infiltration of MCs and Th17 cells) for LUAD prognosis, and, more importantly, establishes a solid theoretical foundation for developing novel immunotherapeutic combination strategies targeting this axis.

Supplementary Information
Below is the link to the electronic supplementary material.