The research team, led by investigators affiliated with Nanchang University, began by asking a deceptively simple question: do ovarian tumours harbour fungi, and if so, does the fungal community differ between benign borderline tumours and full-blown epithelial ovarian cancer? To answer it, they recruited 40 women aged 18 to 75, half with epithelial ovarian cancer (EOC) and half with epithelial borderline ovarian tumours (EBOT), a precursor-like condition that lacks destructive stromal invasion. Tissue was collected under strictly sterile conditions during surgery, snap-frozen in glycerol preservation solution, and subjected to sequencing of the internal transcribed spacer 1 (ITS1) region of fungal ribosomal DNA, the standard barcoding tool for fungal community profiling.

The results were striking. Fluorescence in situ hybridisation using a pan-fungal probe targeting 28S rRNA confirmed that fungal signals were present in both tumour types, but the signals were markedly stronger in the malignant tissue. Sequencing revealed that EOC samples had significantly reduced fungal alpha-diversity, with lower Chao1 indices and fewer observed species, alongside clear separation in beta-diversity clustering. At the genus level, the malignant tumours were significantly enriched in Malassezia and Meyerozyma, while Cutaneotrichosporon was depleted. Quantitative PCR validated the Malassezia enrichment, and, perhaps most provocatively, the abundance of the genus correlated positively with serum CA125, the clinical biomarker used to track ovarian cancer burden. The implication was hard to escape: the more aggressive the disease, the more Malassezia was present.

Correlation, of course, is not causation, and this is where the study distinguishes itself from the many catalogue-style microbiome papers that have accumulated in recent years. To test whether the tumour-associated fungi were doing anything, the team transplanted microbial communities extracted from patient tissues directly into established tumours in a murine model of EOC, using subcutaneously implanted ID8 ovarian cancer cells in C57BL/6 mice. Before transplantation, mice were depleted of their endogenous fungi with amphotericin B, delivered by gavage and in drinking water, to create a clean experimental backdrop.

The outcome was unambiguous. Both EBOT-derived and EOC-derived microbiota accelerated tumour growth, but the EOC-derived communities were markedly more potent. Tumours in the transplant groups grew faster, weighed more, and showed denser populations of atypical neoplastic cells, along with a higher proportion of Ki-67-positive proliferating cells on immunohistochemistry. Critically, body weights remained stable across all groups, suggesting the effect was localised to the tumour microenvironment rather than a systemic toxic effect. The fungal-depleted mice without transplants grew tumours at rates indistinguishable from controls, ruling out amphotericin B itself as a confounder.

With the transplantation model established, the researchers narrowed their focus to Malassezia restricta, one of the most prevalent fungal species previously identified across diverse human tumours. Injecting M. restricta alone directly into tumours reproduced the pro-tumour phenotype almost exactly, matching the effect of the full EOC-derived microbial community. This single fungal species, it turned out, carried most of the tumour-promoting punch.

But how does a yeast manipulate a tumour from within? The team turned to the immune landscape of the tumours, characterising macrophage subsets by immunofluorescence, quantitative PCR, western blot and ELISA. Macrophages in tumours are notoriously plastic, existing along a spectrum from classically activated M1 cells, which produce inflammatory cytokines like TNF-α and IL-6 and fight tumours, to alternatively activated M2 cells, which secrete immunosuppressive IL-10 and TGF-β and express markers such as CD206 and arginase-1 (Arg1). In the M. restricta-colonised tumours, the balance tipped decisively toward M2: CD206-positive cells infiltrated in far greater numbers, CD206 and Arg1 transcripts surged, Arg1 protein rose, and IL-10 and TGF-β climbed. M1 markers, CD4-positive and CD8-positive T-cell populations all remained essentially unchanged. The fungus, in other words, was not broadly perturbing immunity but performing a precise surgical reprogramming of the macrophage compartment.

The search for the molecular mediator led to extracellular vesicles, nanoscale bilayer-membrane particles that microbes and host cells alike use as intercellular couriers, ferrying proteins, nucleic acids, lipids and metabolites between cells. When the researchers grew M. restricta in culture and separated its supernatant from its heat-killed cell bodies, only the supernatant induced M2 polarisation in RAW264.7 macrophages, pointing toward a soluble, secreted effector. Ultracentrifugation of the supernatant at 100,000 × g yielded a purified vesicle fraction. Transmission electron microscopy revealed the classic cup-shaped morphology of extracellular vesicles, and nanoparticle tracking analysis placed their diameter between 50 and 300 nanometres, averaging 154.8 nanometres. Labelling the vesicles with the fluorescent dye PKH26 showed that macrophages avidly internalised them within 24 hours, with the vesicles accumulating in the perinuclear region.

Once inside, the vesicles got to work. Treating macrophages with the purified M. restricta extracellular vesicles (MrEVs) at 20 micrograms per millilitre reproduced the full M2 signature: elevated CD206 and Arg1, increased IL-10 and TGF-β secretion, and the phenotypic shift confirmed by multiple independent assays. Screening candidate signalling pathways revealed that the vesicles selectively upregulated TLR4, the Toll-like receptor 4 that serves as a key sensor of microbial ligands, and JAK2, the Janus kinase at the head of one of the most consequential inflammatory cascades in immunology. Western blotting confirmed the downstream consequences: increased phosphorylation of both JAK2 and its transcription factor target STAT3, the canonical molecular switch that drives macrophages toward the M2 fate.

The mechanistic proof came through pharmacological inhibition. AG490, a selective JAK2 inhibitor, was applied both to MrEV-treated macrophages in vitro and to M. restricta-colonised mice in vivo. In the cell experiments, AG490 suppressed the vesicle-induced phosphorylation of JAK2 and STAT3, prevented the rise in CD206 and Arg1, blocked IL-10 and TGF-β secretion, and critically, eliminated the ability of MrEV-conditioned macrophages to stimulate the proliferation and migration of ID8 ovarian cancer cells in wound-healing and CCK-8 assays. In the mice, AG490 administration alongside M. restricta colonisation shrank tumours, reduced Ki-67 staining, blocked CD206-positive M2 infiltration and dampened Arg1 and JAK2/STAT3 phosphorylation in the tumour tissue. The pathway, in short, was both necessary and sufficient for the fungal effect.

The significance of this work extends well beyond ovarian cancer. Over the past several years, the intratumoural microbiome has transitioned from a contested idea to an established field, with bacterial communities documented in pancreatic, breast, lung and colorectal malignancies, and fungal communities mapped across 35 cancer types. Earlier studies had already implicated Malassezia in other cancers: M. globosa accumulation in breast tumours drives IL-17A-dependent M2 polarisation, and Malassezia enrichment in hepatocellular carcinoma correlates with poor prognosis by reshaping bile acid metabolism. But the present study is among the first to trace a complete causal chain in ovarian cancer, from fungal presence in patient tissue, through transplantation and mono-colonisation in living animals, to a defined molecular effector, a defined receptor-pathway axis, and a pharmacological rescue.

It also opens a genuinely new therapeutic door. The JAK2/STAT3 pathway is already druggable, with inhibitors in clinical use for myelofibrosis and under investigation across oncology, and the tumour-associated macrophage phenotype it controls is a validated target of immuno-oncology programs worldwide. If M. restricta enrichment and its vesicle-mediated signalling can be confirmed in larger patient cohorts, measures of intratumoural fungal load might serve as biomarkers for selecting patients likely to benefit from JAK-pathway modulation or macrophage-reprogramming strategies. Conversely, interventions that displace or suppress pro-tumourigenic fungi, as amphotericin B did in the mouse model, could complement existing treatments, though the authors caution that much work remains.

Limitations deserve honest mention. The clinical cohort was modest, 20 patients per group, and the colonisation burden of Malassezia was not quantitatively measured with digital PCR or metagenomic tools. Germ-free mouse models would eliminate residual confounding from endogenous fungi, and the specific molecular cargo within MrEVs that engages TLR4 remains unidentified. The authors suggest proteomic and small RNA sequencing approaches to pinpoint these ligands, and note that complete JAK2 knockout models would strengthen the pathway validation beyond what AG490’s partial inhibition can achieve.

Still, the central message stands. A fungus best known for contributing to dandruff and seborrheic dermatitis has been shown, through a rigorous chain of patient sequencing, animal transplantation, vesicle purification and pathway rescue, to help ovarian cancer grow by hijacking the very immune cells meant to fight it. It is a vivid reminder that tumours are not merely masses of mutated cells but complex ecosystems, and that within those ecosystems, even the smallest and most overlooked inhabitants may hold the keys to new diagnostics and new cures. The fungal-immune axis in cancer, once invisible, is now in plain sight, and it is unlikely to be ignored again.

For decades, the focus in CRC has centered on genetic mutations and signaling aberrations within tumor cells themselves. However, the dense network of immune populations surrounding these malignant cells profoundly influences whether a tumor grows aggressively or succumbs to immune control. Among these immune actors, macrophages—a versatile type of white blood cell—play a dichotomous role. When polarized into the M1 phenotype, macrophages attack tumor cells and promote inflammatory responses. Conversely, the M2-polarized macrophages foster tissue repair but, regrettably, also create an immunosuppressive niche that supports tumor growth, invasion, and metastasis. Understanding what switches macrophages into this tumor-supportive M2 state is crucial for innovative immunotherapies.

The recent research sheds light on MIIP as a master regulator restraining macrophages from adopting the pro-tumor M2 phenotype. Their comprehensive approach combined transcriptomic analyses of large public colorectal cancer datasets, mechanistic studies in cell cultures, and validation in animal models. Intriguingly, they revealed a suppressive loop where loss of MIIP triggers cytoplasmic DNA stress, setting off the stimulator of interferon genes (STING) pathway—a critical innate immune sensor of aberrant DNA. Activation of STING physiologically orchestrates immune responses against pathogens and damaged cells. Yet, in the tumor context, chronic STING activation fuels a detrimental cascade involving the non-canonical NFκB2 signaling pathway, culminating in increased secretion of IL-10, an immunosuppressive cytokine that tilts macrophages towards the M2 phenotype.

This molecular interplay between MIIP repression, STING activation, NFκB2 signaling, and IL-10 production establishes a feedback loop wherein colorectal cancer cells effectively co-opt macrophages to become their unwitting allies. Experimental reduction of MIIP in cancer cell lines induced STING pathway upregulation, elevated NFκB2-driven IL-10 secretion, and resulted in pronounced M2 macrophage polarization when these cells were co-cultured. Macrophages conditioned in this microenvironment exhibited gene expression profiles and surface markers characteristic of M2-type cells, known to facilitate tumor invasion and metastasis.

Crucially, the study provided compelling in vivo evidence using murine models of colorectal cancer. Tumors engineered to express high MIIP levels showed slower growth rates, markedly fewer liver metastases, and reduced infiltration by M2 macrophages compared to MIIP-deficient counterparts. Pharmacological blockade of STING reversed these effects, reinforcing the therapeutic potential of targeting this axis. Analyses of clinical CRC samples corroborated the inverse relationship between MIIP expression and STING activity, IL-10 levels, and the abundance of M2 macrophages, making MIIP a promising biomarker predicting tumor immune landscape and patient prognosis.

The revelations brought forth challenge the traditional view of MIIP solely as a cell-intrinsic tumor suppressor. Instead, MIIP emerges as a critical regulator of tumor-immune crosstalk, dictating whether the microenvironment fosters immune destruction or immune evasion of cancer cells. By modulating macrophage polarization, MIIP influences not only tumor growth but also metastatic potential and immune resistance, phenomena that have long frustrated clinicians and researchers alike.

This study’s deeper insight into the STING-NFκB2-IL10 signaling axis illuminates a novel immune regulatory mechanism exploited by CRC cells. Unlike classical inflammatory NFκB signaling that promotes anti-tumor immunity, non-canonical NFκB2 activation here culminates in immune suppression via IL-10, enabling a pro-cancer environment. Importantly, the persistent activation of STING signaling by DNA stress in MIIP-deficient tumors contrasts with its canonical role in anti-viral defense, revealing how tumors repurpose innate immune pathways to their advantage.

From a therapeutic standpoint, these findings invigorate interest in modulating the tumor microenvironment by targeting macrophage polarization and STING signaling. Given that most colorectal cancers remain “immune-cold”—lacking robust immune infiltration and resisting checkpoint inhibitors—strategies that disrupt the MIIP-STING axis could reactivate effective anti-tumor immunity. This paradigm shift advocates for combinatorial therapies integrating immune microenvironment reprogramming agents with existing immunotherapies to overcome resistance and reduce metastasis.

Moreover, the study suggests measuring MIIP levels in tumors could stratify patients most likely to benefit from STING-targeted agents or macrophage-modulating drugs. Personalized immunotherapy approaches grounded in tumor microenvironment profiling may thus represent the next frontier in improving CRC patient outcomes. As immune checkpoint inhibitors alone fall short in the majority of CRC cases, such integrative strategies provide renewed hope in transforming treatment landscapes.

Beyond colorectal cancer, the mechanistic insights into MIIP’s role highlight a broad principle of tumor progression control: regulating immune cell behavior within the tumor milieu can be as critical as targeting malignant cells directly. Tumors can exploit immune signaling pathways to evade destruction and fuel metastasis, emphasizing the importance of decoding these complex dialogues for future drug development.

In sum, this pioneering investigation charts new territory at the intersection of chromosomal instability, immune signaling, and tumor progression. By elucidating how MIIP controls macrophage polarization via the STING–NFκB2–IL10 axis, the research not only deepens our understanding of colorectal cancer biology but also opens avenues for innovative precision immunotherapies that reshape the immune landscape to fight cancer more effectively.

Subject of Research:
Article Title: Migration and invasion inhibitory protein inhibits M2 macrophage polarization to suppress colorectal cancer progression through the STING–NFκB2–IL10 axis
News Publication Date: 14-Jan-2026
References: 10.20892/j.issn.2095-3941.2025.0282
Image Credits: Cancer Biology & Medicine
Keywords: Colorectal cancer, tumor microenvironment, macrophage polarization, migration and invasion inhibitory protein, MIIP, STING pathway, NFκB2 signaling, IL-10 cytokine, immune suppression, immunotherapy, tumor progression, cancer metastasis