The study began with a systematic survey. Rather than focusing on a single candidate gene, the researchers profiled the entire serpin family of genes in head and neck squamous cell carcinoma using data from The Cancer Genome Atlas, the large-scale international effort that has catalogued genomic and clinical information from thousands of tumor specimens. Among the members of this family, best known for proteins that regulate protease activity in blood clotting and tissue remodeling, one gene stood out: SERPINE1, which encodes plasminogen activator inhibitor-1, a secreted protein previously implicated in tumor invasion and metastatic spread in multiple cancer types. Elevated SERPINE1 expression correlated with poor outcomes, making it the primary gene of interest for the team led by corresponding author Dapeng Lei.

To move from correlation to causation, the researchers built laboratory models in which SERPINE1 was either stably silenced or artificially overexpressed in head and neck cancer cell lines. The functional consequences were unambiguous. Cells stripped of SERPINE1 proliferated more slowly, migrated less efficiently across wound surfaces, and showed a markedly reduced capacity to invade through extracellular matrix, the biological hallmark that distinguishes an in situ lesion from a truly dangerous carcinoma. Conversely, cells engineered to produce excess SERPINE1 displayed enhanced versions of all three malignant behaviors. The team then extended these in vitro findings into living animals by transplanting the manipulated cells subcutaneously into immunodeficient nude mice, where the growth differences between SERPINE1-high and SERPINE1-low tumors could be observed directly.

With the oncogenic role of SERPINE1 established, the central question became mechanistic: through which downstream genes does this protein exert its effects? The researchers turned to transcriptome sequencing, an unbiased approach that measures the activity of every gene in the genome simultaneously. Comparing the transcriptional landscapes of SERPINE1-depleted cells against controls, they identified carbonic anhydrase 9, abbreviated CA9, as a key effector. CA9 encodes a membrane-spanning enzyme that catalyzes the reversible hydration of carbon dioxide to bicarbonate and protons, a reaction central to pH regulation. The enzyme is famously hypoxia-inducible, driven by HIF-1 transcriptional activity in oxygen-starved tumor regions, and it helps cancer cells maintain an alkaline intracellular environment while acidifying their surroundings, a shift that promotes invasion and blunts the efficacy of weakly alkaline chemotherapeutics.

The connection between SERPINE1 and CA9 proved especially consequential in the context of cisplatin treatment. To study drug resistance rigorously, the team generated cisplatin-resistant derivatives of two head and neck cancer cell lines, designated FaDu-CDDP and Tu686-CDDP, by exposing the parental cells to escalating doses of the drug. In these resistant populations, both SERPINE1 and CA9 were found to be upregulated in tandem, suggesting that the resistant phenotype travels along the same regulatory axis identified in the sequencing data. Quantitative measurements of the half-maximal inhibitory concentration, or IC50, confirmed that the resistant lines required substantially more cisplatin to achieve the same degree of cell killing, validating the model system before any genetic perturbations were attempted.

The rescue experiments that followed constitute the most technically decisive portion of the study. When the researchers silenced SERPINE1 in the resistant cells, the cells regained sensitivity to cisplatin, and silencing CA9 produced the same effect, indicating that either node of the axis could be targeted to weaken drug resistance. The critical test came next: in SERPINE1-depleted cells, the team restored CA9 expression. Remarkably, this single intervention reversed the suppressive phenotypes caused by SERPINE1 loss, reinstating both the aggressive growth characteristics and, crucially, the resistance to cisplatin. In the logic of molecular biology, such rescue experiments are the gold standard for demonstrating that a downstream gene is not merely a correlate but a functional mediator, and the results placed CA9 squarely in that role, operating downstream of SERPINE1 to execute its effects on tumor behavior and drug response.

The findings carry significant weight for a cancer type in which effective molecular targets remain frustratingly scarce. Unlike lung cancer, where EGFR mutations and ALK rearrangements have spawned a generation of targeted therapies, head and neck squamous cell carcinoma has seen relatively few successes in precision oncology, with anti-PD-1 immunotherapy representing one of the few meaningful advances in recent years. A biomarker axis that simultaneously predicts prognosis and dictates sensitivity to a widely used chemotherapy drug would fill a genuine clinical void. If SERPINE1 and CA9 expression levels can stratify patients at diagnosis, oncologists might one day identify which individuals are likely to benefit from cisplatin and which should be routed toward alternative regimens or clinical trials of combination therapies designed to disrupt the axis itself.

There are also therapeutic implications embedded in the biology of the two proteins. CA9 has long attracted pharmaceutical interest because of its accessibility on the cell surface and its well-characterized enzymatic active site, and several carbonic anhydrase IX inhibitors have progressed through preclinical development, some investigated in combination with standard chemoradiation. The new data suggest that such inhibitors could be particularly valuable in tumors with high SERPINE1 expression, where CA9 acts as the functional conduit for both malignancy and cisplatin resistance. SERPINE1 itself, as a secreted serine protease inhibitor, presents a different but not insurmountable drug development challenge, and small-molecule and biologic approaches to blocking plasminogen activator inhibitor-1 have been explored in other disease contexts, including fibrosis and cardiovascular pathology.

The study also illustrates the power of a multi-layered research design that moves from population-level genomics through engineered cell models to animal xenografts and mechanistic rescue experiments. By anchoring their analysis in TCGA data, the researchers ensured that their candidate gene was relevant to real patient tumors rather than an artifact of cell culture. By deploying two independent cisplatin-resistant cell lines, FaDu and Tu686, they guarded against conclusions that might hold true for only a single genetic background. The work received funding from the National Natural Science Foundation of China and provincial science programs in Shandong, and it was approved by the institutional ethics committees for both the human tissue analyses and the animal procedures, with the authors declaring no competing interests.

Caveats remain, as they always do in translational cancer research. The mechanistic chain was established in cell lines and mouse xenografts, and the abstract does not report clinical trial data demonstrating that SERPINE1 or CA9 levels predict cisplatin response in actual patient cohorts, an essential step before any biomarker enters routine practice. The precise molecular circuitry connecting SERPINE1, a secreted protease inhibitor, to CA9, a hypoxia-responsive metabolic enzyme, also invites further dissection, since the intermediate signaling steps likely involve pathways at the interface of tumor metabolism and stress adaptation. Nevertheless, the demonstration that a single axis governs both tumor progression and chemotherapy resistance in head and neck squamous cell carcinoma offers a coherent framework for future studies, and it revives attention to carbonic anhydrase IX as a vulnerability that may be exploitable precisely at the moment when cisplatin’s effectiveness is faltering. For patients facing a diagnosis with an unfavorable prognosis and limited targeted options, that convergence of growth and resistance in one druggable pathway is a lead worth chasing.

Subject of Research: The SERPINE1–CA9 axis driving progression and cisplatin resistance in head and neck squamous cell carcinoma

Article Title: SERPINE1 promotes HNSCC progression and cisplatin resistance by upregulating CA9

Article References: Wang, S., Chang, F., Peng, T., Li, Z., Wang, Y., Duan, C., Wu, K., Wang, H., Wei, D., Li, W., Qian, Y., Cao, S., Wang, Y., Liu, J., Fang, J., Chen, L., Zhao, J., & Lei, D. (2026). SERPINE1 promotes HNSCC progression and cisplatin resistance by upregulating CA9. Cellular and Molecular Life Sciences. https://doi.org/10.1007/s00018-026-06439-z

Image Credits: AI Generated

DOI: 10.1007/s00018-026-06439-z

Keywords: HNSCC, SERPINE1, CA9, carbonic anhydrase IX, cisplatin resistance, tumor progression, drug resistance, TCGA, xenograft, hypoxia, biomarker, molecular oncology

The review, published on 3 September 2026, does not claim that radiomics should replace tissue-based biomarkers. Its central argument is more subtle and, arguably, more clinically useful: radiomics should serve as a spatial context that tells clinicians when a biopsy result can be trusted and when it might be dangerously misleading. Head and neck squamous cell carcinoma, or HNSCC, is an ideal testing ground for this idea. These tumors arise in anatomically complex regions — the larynx, oropharynx, oral cavity, hypopharynx — where treatment decisions carry enormous consequences for speech, swallowing, and quality of life, and where the stakes of both overtreatment and undertreatment are exceptionally high.

To understand why spatial context matters, it helps to distinguish between biomarkers that vary across a tumor and those that do not. The authors organize tissue-based biomarkers along a spectrum of spatial stability. HPV and p16 status, which define a biologically distinct and generally more treatable subtype of oropharyngeal cancer, tend to be uniform throughout the tumor, so a single biopsy usually captures them reliably. Certain genomic alterations behave similarly. But other clinically decisive biomarkers are far less cooperative. PD-L1 expression, the density and location of immune-cell infiltration, necrosis, and immune exclusion — the phenomenon where immune cells are kept at the tumor’s periphery — can all vary dramatically from one region of a tumor to another. A core needle sample taken from a well-perfused, immune-inflamed edge may paint a completely different picture than a hypoxic, necrotic core just a few millimeters away.

This is not a theoretical concern. The landmark multiregion sequencing work of Gerlinger and colleagues, published in the New England Journal of Medicine in 2012, demonstrated that intratumor heterogeneity and branched evolution are fundamental features of cancer biology. In HNSCC, the clinical consequences are concrete. When a tumor board decides whether a patient with HPV-positive oropharyngeal cancer can safely receive reduced-intensity treatment — so-called de-escalation — the decision often hinges on biomarkers measured in a single sample. If that sample happens to come from an unrepresentative region, the de-escalation decision rests on sand. Similarly, eligibility for immunotherapy frequently depends on PD-L1 scoring, and the review points out that localized sampling may incompletely capture heterogeneity in exactly these selected clinical contexts.

Radiomics offers a fundamentally different vantage point. The technique extracts hundreds of quantitative features from routinely acquired medical images — computed tomography, magnetic resonance imaging, and FDG-PET — converting the visual texture, shape, and intensity patterns of a tumor into high-dimensional data. First articulated in a widely cited 2014 Nature Communications paper by Aerts and colleagues, the premise is that image features reflect underlying biological processes: coarse heterogeneity on CT may mirror necrosis and hypoxia; specific texture patterns on diffusion-weighted MRI, measured through the apparent diffusion coefficient, correlate with cell density and oxygenation; FDG-PET textural features track metabolic activity and its spatial distribution. Because imaging captures the entire tumor volume — and often the surrounding peritumoral tissue — it sees what the biopsy cannot.

The review synthesizes a substantial body of evidence linking radiomic features to the biological processes that matter most in HNSCC. Tumor hypoxia is a classic example: it has been known since Brizel and colleagues’ 1997 work to worsen prognosis in head and neck cancer, largely because oxygen-deprived cells resist radiation. Hypoxia is also profoundly spatial — it develops in regions distant from functional blood vessels — which makes it a natural target for imaging. Studies have shown that textural features of hypoxia PET predict survival during chemoradiotherapy, and recent work on apparent diffusion coefficient MRI suggests it can act as a predictive biomarker for hypoxia, treatment de-escalation, and recurrence in HPV-associated oropharyngeal cancer. On the immune side, radiomic signatures have been associated with PD-L1 expression, CD8-positive T-cell infiltration, T cell-inflamed gene expression profiles, and even gamma-delta T-cell abundance, while radiogenomic analyses have connected imaging heterogeneity to somatic mutations in genes such as TP53, FAT1, and KMT2D.

What sets the Keio review apart from much of the radiomics literature is its insistence on mapping these features to specific clinical decision points rather than treating them as generic prognostic scores. The authors walk through three major scenarios. The first is larynx preservation, where the choice between organ-sparing chemoradiotherapy — a strategy established by the landmark 2003 trial of concurrent chemotherapy and radiotherapy — and primary surgery depends on predicting response. Radiomic features reflecting hypoxia and necrosis could flag tumors whose biopsy-based profiles look favorable but whose whole-tumor biology suggests resistance. The second is immunotherapy stratification, where combining a PD-L1 score from a biopsy with imaging evidence of immune exclusion or spatially restricted immune activity could refine patient selection for checkpoint inhibitors such as pembrolizumab and nivolumab, agents whose benefit in recurrent or metastatic HNSCC was established in the KEYNOTE-048 and CheckMate 141 trials. The third is recurrence assessment, where peritumoral radiomic signatures and subregion-based models have shown promise in predicting locoregional failure after chemoradiotherapy and even the site of recurrence after reirradiation.

The review also embraces a dynamic dimension of imaging that tissue sampling cannot match: delta radiomics, the analysis of how radiomic features change over time. Serial multiparametric MRI and FDG-PET acquired during the course of radiation therapy have been shown to predict treatment response, and FDG-PET can identify pathological response early during neoadjuvant immune checkpoint blockade. In principle, a clinician could watch the spatial biology of a tumor evolve week by week during treatment and adjust course accordingly — something no biopsy workflow could realistically achieve. Combined with deep-learning segmentation tools that automate tumor delineation on CT and MRI, and with harmonization methods such as ComBat that correct for differences between scanners and institutions, the technical pipeline for deploying these approaches at scale is maturing rapidly.

The authors are careful about the limits of the evidence. Nearly all of the studies they synthesize are retrospective, and the field has long struggled with reproducibility: radiomic features can be sensitive to image acquisition parameters, reconstruction algorithms, and segmentation choices, which is why standardization initiatives such as the Image Biomarker Standardisation Initiative have become essential infrastructure. Multi-institutional modeling studies have begun to evaluate how well radiomics and deep-learning models generalize beyond their development cohorts, and systematic reviews of machine-learning models for predicting radiation toxicity have taken a sober look at actual versus claimed performance. The review’s framing — radiomics as context rather than standalone predictor — is partly a response to this uncertainty. A model that claims to replace PD-L1 testing must clear an enormous evidentiary bar; a tool that flags when a biopsy result deserves a second look needs to be merely reliable enough to prompt caution.

That pragmatic framing may prove to be the review’s most lasting contribution. Rather than promising a revolution, Kono and colleagues describe an incremental integration: radiomics as a layer of spatial intelligence woven into existing biomarker-guided workflows, identifying the patients in which hypoxia, necrosis, stromal architecture, or immune exclusion make a small sample untrustworthy. For a disease where the difference between de-escalation and full-dose treatment can mean preserving a voice, and where immunotherapy decisions hinge on biomarkers that flicker across the tumor landscape, that kind of whole-tumor perspective could change how multidisciplinary teams weigh the evidence in front of them. The next step, the authors imply, is prospective validation — testing whether spatially informed caution actually improves outcomes. If it does, the humble biopsy may finally gain the companion it has always needed: a map of everything it cannot see.

Subject of Research: Use of radiomic imaging features as spatial context for interpreting biopsy biomarkers in head and neck squamous cell carcinoma treatment decisions

Article Title: Radiomics as a spatial context for treatment decision-making in head and neck cancer

Article References: Kono, T., Kasahara, K., Ogawa, R., & Ozawa, H. (2026). Radiomics as a spatial context for treatment decision-making in head and neck cancer. British Journal of Cancer. https://doi.org/10.1038/s41416-026-03600-0

Image Credits: AI Generated

DOI: 10.1038/s41416-026-03600-0

Keywords: radiomics, head and neck cancer, HNSCC, spatial heterogeneity, PD-L1, HPV, hypoxia, immunotherapy, biomarkers, de-escalation, medical imaging, tumor microenvironment

The pathway in question is the hepatocyte growth factor (HGF)/c-MET axis, a receptor tyrosine kinase system that normally orchestrates embryonic development, tissue regeneration, and wound healing. In cancer, it is hijacked. Overexpression of c-MET has been documented in more than eighty percent of HNSCC cases and correlates with poor prognosis, and the pathway is implicated in resistance to EGFR-targeted drugs such as cetuximab. Earlier work from the same group had already shown that HGF stimulation ramps up glycolysis—the glucose-fermenting Warburg metabolism familiar from textbook cancer biology—in head and neck cancer cells. What remained unknown was whether the same signal also governs glutamine metabolism, the second great nutrient stream on which proliferating tumors depend.

Glutamine is far more than a backup fuel. For rapidly dividing cells it supplies both carbon and nitrogen for building nucleotides, amino acids, and lipids. The gatekeeper of this economy is glutaminase 1, or GLS-1, the rate-limiting enzyme that converts glutamine into glutamate, thereby feeding the tricarboxylic acid cycle and powering mitochondrial energy production. Elevated GLS-1 has been tied to aggressive behavior and poor outcomes in multiple cancer types. The Würzburg-led team, including first author Marius Hörner and senior author Stefan Hartmann, reasoned that if c-MET could commandeer glucose handling, it might also pull the levers of glutaminolysis—and that the connection could be exploitable.

To find out, the researchers compared basal GLS-1 levels in primary human oral keratinocytes—the non-malignant controls—with three HNSCC cell lines of distinct MET status: MET-amplified Detroit 562, originally derived from a pleural metastasis of a pharyngeal carcinoma, and the MET wild-type lines FaDu and SCC-154. The result was unambiguous. Every cancer line expressed more GLS-1 than normal keratinocytes, and the MET-amplified Detroit 562 cells showed the strongest basal expression. When the cells were bathed in recombinant HGF for twenty-four hours, only Detroit 562 responded dramatically: GLS-1 mRNA surged approximately 3.4-fold above control, and GLS-1 protein climbed several-fold on Western blot. FaDu and SCC-154 cells, lacking MET amplification, barely blinked. The metabolic reprogramming, in other words, was selective for cells with an amplified receptor.

The selectivity proved mechanistically decisive. Pretreating Detroit 562 cells with Foretinib, a selective c-MET inhibitor, completely abolished the HGF-driven rise in GLS-1, restoring expression to baseline, while the inhibitor alone left basal levels untouched. Functional assays tracked the metabolic consequences in the culture medium itself. After HGF stimulation, extracellular glutamine concentrations dropped significantly—evidence of increased consumption—while extracellular glutamate rose sharply, consistent with accelerated glutaminolysis. Foretinib prevented both changes, pinning the flux shift squarely on c-MET activity. For a cell line already prone to metastatic behavior, the growth factor was effectively upgrading its metabolic engine.

Next came the question of which intracellular cascade transmits the signal. Based on the group’s prior glucose-metabolism work, in which neither PI3K/AKT nor JAK/STAT contributed to the metabolic switch, attention turned to MAPK/ERK. The bet paid off. HGF triggered robust phosphorylation of ERK1/2 at Tyr204 in Detroit 562 cells, and the selective ERK1/2 inhibitor FR180204 completely blocked the HGF-induced increase in GLS-1 at both mRNA and protein levels. The signaling logic was thus clean: HGF binds c-MET, c-MET activates ERK, and ERK drives GLS-1 transcription—an oncogenic kinase cascade wired directly into a metabolic enzyme.

The functional fallout was equally clear. In wound-healing assays, HGF-stimulated Detroit 562 monolayers closed scratches at 65.4 percent after twenty-four hours and 89.7 percent after forty-eight hours, compared with 40.3 and 70.5 percent in untreated controls—a marked boost in motility. ERK inhibition erased this pro-migratory effect, and HGF also increased phosphorylation of focal adhesion kinase, a surrogate marker of migratory activity, in a c-MET-dependent fashion. When the researchers silenced GLS-1 with siRNA, wound closure slowed significantly, demonstrating that the enzyme is not a passive bystander but a functional requirement for HGF-driven migration. Cell proliferation markers fell in parallel.

Mitochondrial physiology told the same story with quantitative precision. Using Seahorse extracellular flux analysis, which tracks the oxygen consumption rate (OCR) of living cells through sequential injection of oligomycin, the uncoupler BAM15, and rotenone/antimycin A, the team showed that HGF significantly raised both basal and maximal mitochondrial respiration in MET-amplified Detroit 562 cells. GLS-1 knockdown dropped basal and maximal OCR and abolished the HGF-induced respiratory boost. In MET wild-type FaDu cells, by contrast, HGF left respiration unchanged, and GLS-1 silencing reduced OCR independently of HGF—confirming that the metabolic response is a specific property of the MET-amplified state, not a general feature of head and neck cancer cells.

Critically, the laboratory findings survived contact with real patient data. Mining RNA-sequencing data from 522 primary tumors in the TCGA-HNSC cohort, the researchers found a significant positive Spearman correlation between MET and GLS1 expression (ρ = 0.432, p < 0.0001), a weaker but significant link between HGF and GLS1 (ρ = 0.165), and the strongest correlation of all between a composite MAPK/ERK transcriptional activity score—built from ten established ERK target genes including DUSP6, MYC, and CCND1—and GLS1 expression (ρ = 0.581, p < 0.0001). Stratifying patients by median MET expression, MET-high tumors showed significantly higher GLS1, along with elevated SLC1A5, the glutamine transporter, whereas the related gene GLS2 did not differ—underscoring the specificity of GLS1-dependent glutaminolysis downstream of c-MET.

The therapeutic implications are tantalizing, though the authors are careful to frame them appropriately. Single-agent c-MET inhibitors such as Foretinib and Tivantinib have shown limited efficacy in HNSCC, but the new data suggest that dual targeting of c-MET and GLS-1 could circumvent compensatory signaling, and next-generation glutaminase inhibitors such as CB-839 have already demonstrated activity in other malignancies. The bispecific EGFR/MET antibody Amivantamab, approved in non-small cell lung cancer and now showing promising early results in the OrigAMI-4 trial for HNSCC patients who progressed on checkpoint inhibitors and chemotherapy, adds further momentum. Because c-MET signaling has also been linked to PD-L1 induction and glutamine availability shapes antitumor immunity, combining MET blockade, GLS-1 inhibition, and immunotherapy is an openly debated possibility. Caveats remain: the study relied on 2D culture models centered on a single MET-amplified line, intracellular isotope-tracing flux measurements were not performed, and in vivo efficacy was not tested. Validation in 3D spheroids, patient-derived organoids, and animal models will be essential. Still, the work expands c-MET’s portfolio beyond proliferation and migration into metabolic sovereignty—and in doing so, hands oncologists a second target in a cancer that desperately needs more than one.

Subject of Research: HGF/c-MET signaling drives glutamine metabolism via MAPK/ERK-dependent GLS-1 induction in MET-amplified head and neck squamous cell carcinoma

Article Title: HGF/c-MET signaling induces glutamine metabolism in MET-amplified head and neck squamous cell carcinoma via MAPK/ERK-dependent induction of GLS-1

Article References: HGF/c-MET signaling induces glutamine metabolism in MET-amplified head and neck squamous cell carcinoma via MAPK/ERK-dependent induction of GLS-1. (n.d.). https://doi.org/10.1007/s12032-026-03409-0

Image Credits: AI Generated

DOI: 10.1007/s12032-026-03409-0

Keywords: HNSCC, c-MET, HGF, GLS-1, glutamine metabolism, MAPK/ERK, Foretinib, glutaminolysis, TCGA, cancer metabolism, mitochondrial respiration, Head and neck cancer