NDMA is one of the most extensively studied nitrosamine carcinogens. It appears in preserved and smoked foods, in contaminated drinking water, and it made international headlines when it was found in batches of the blood pressure drug valsartan and the heartburn medication ranitidine. The liver is its principal target organ, and prolonged low-dose exposure is known to predispose both animals and humans to hepatocellular carcinoma, the most common form of primary liver cancer, which accounts for roughly ninety percent of such cases and remains a leading cause of cancer mortality worldwide. Yet the question of whether NDMA does more than damage DNA — whether it can actively generate the self-renewing, treatment-resistant cells that make liver tumors so hard to eradicate — had remained unresolved.

To address this, the team, led by researchers at Nanjing Medical University, Ningbo No. 2 Hospital, and Sun Yat-sen University Cancer Center, exposed the non-cancerous human liver cell line LO2 to low concentrations of NDMA continuously for thirty passages, a period of about three months. The cells that received the higher dose, one microgram per milliliter, transformed dramatically. They became irregular and spindle-shaped, proliferated faster, migrated and invaded more aggressively than even the established liver cancer line Huh-7 used as a positive control, and showed elevated levels of hepatocellular carcinoma markers such as AFP, GGT1, and GSTP1 alongside a shift toward mesenchymal proteins like vimentin and N-cadherin, a hallmark of epithelial-mesenchymal transition. The researchers designated these cells TLO2 and used them as their experimental model of NDMA-driven liver cancer.

When the transformed cells were cultured in serum-free medium — a standard technique for enriching cancer stem cells — they formed large, abundant spherical clusters, a sign of self-renewal capacity. Flow cytometry revealed that roughly a quarter more cells expressed CD133, a well-established liver cancer stem cell marker, after this enrichment. Levels of additional stemness markers, including SOX2, OCT4, and EPCAM, rose in parallel. Soft agar and colony formation assays confirmed that the NDMA-transformed cells could grow without a solid substrate and form colonies at rates comparable to mature liver cancer cells. In short, chronic NDMA exposure did not merely convert liver cells into tumor-like cells; it endowed them with the stem-like properties that fuel metastasis, recurrence, and drug resistance.

The study then turned to ZNF652, a zinc finger transcription factor that previous work had linked to tumor suppression in several cancers. Database analyses using KM-PLOTTER and TNMplot showed that higher ZNF652 expression correlated with better survival in hepatocellular carcinoma patients, and that ZNF652 levels were lower in metastatic lesions than in primary tumors. In the NDMA-transformed cells, ZNF652 protein dropped after sphere enrichment. When the researchers forced ZNF652 back into the cells, spheres became smaller and less frequent, and the stemness markers CD133, SOX2, OCT4, and EPCAM all declined. Conversely, silencing ZNF652 with siRNA produced larger, more numerous spheres and higher marker expression. These results positioned ZNF652 as a brake on liver cancer stemness.

The Sonic Hedgehog pathway, a classical developmental signaling cascade that is frequently hijacked in cancer, emerged as the mechanism through which ZNF652 acts. Database comparisons showed that key components of the pathway — SHH, SMO, GLI1, and GLI2 — were elevated in liver cancer tissue relative to normal liver, and that they rose further when the transformed cells were enriched for stemness. GLI1, the transcriptional effector that translocates to the nucleus to activate Hedgehog target genes, proved particularly consequential: overexpressing it increased sphere size and frequency and boosted stemness markers, while knocking it down had the opposite effect. The pathway, in other words, actively promotes the cancer stem cell state in these cells.

The pivotal finding came when the team connected the two molecules. Correlation analyses revealed a negative relationship between ZNF652 and the Hedgehog components, and manipulating ZNF652 levels produced reciprocal changes in SHH, SMO, GLI1, and GLI2 expression. Using the JASPAR database, the researchers identified potential ZNF652 binding sites in the GLI1 promoter, then confirmed with dual-luciferase reporter assays that ZNF652 overexpression suppressed GLI1 transcriptional activity. Rescue experiments sealed the logic: when GLI1 was artificially raised, adding ZNF652 brought CD133 expression back down, and when ZNF652 was knocked down, reducing GLI1 reversed the resulting increase in stemness. ZNF652, the data showed, directly targets GLI1 to restrain the Hedgehog-driven cancer stem cell program.

Sulforaphane entered the picture as the pharmacological lever on this axis. The compound, produced when the enzyme myrosinase converts glucoraphanin in broccoli and other cruciferous vegetables, has long been known for anti-inflammatory, antioxidant, and anticancer properties, and ongoing clinical trials have explored its ability to inhibit cancer development. In the new study, treating the transformed cells with sulforaphane at concentrations of one, three, and five micromolar reduced sphere size and frequency, suppressed colony formation, and diminished self-renewal capacity as measured by extreme limiting dilution analysis. Western blots showed dose-dependent declines in CD133, SOX2, OCT4, and EPCAM, and immunofluorescence confirmed reduced CD133 signal. Crucially, sulforaphane also lowered the Hedgehog pathway components and raised ZNF652, and it reversed the stemness increase caused by either GLI1 overexpression or ZNF652 knockdown — evidence that the compound works through the same axis it was suspected of modulating.

To test whether these effects held in living tissue, the researchers implanted NDMA-pretreated Huh-7 cells subcutaneously into twenty female BALB/c nude mice, dividing them into control, NDMA-exposed, and two sulforaphane treatment groups receiving twelve point five or twenty-five milligrams per kilogram by oral gavage twice weekly for four weeks. Tumors in the NDMA group grew significantly larger than controls, while both sulforaphane doses reduced tumor volume and weight in a dose-dependent manner. Immunohistochemistry on excised tumors showed that NDMA exposure elevated CD133 and GLI1 while suppressing ZNF652, and that sulforaphane reversed all three changes, restoring ZNF652 and lowering the stemness and Hedgehog markers.

The authors acknowledge limitations: the xenograft model relied on cells pretreated with NDMA before implantation rather than systemic exposure in the animals, so it may not fully reproduce the natural course of NDMA-associated hepatocarcinogenesis or the complexity of the liver microenvironment, and the human database findings are correlative and were not validated in independent patient cohorts. Even so, the study offers a coherent mechanistic account of how a dietary carcinogen can generate the cellular seeds of liver cancer recurrence, and how a compound found in everyday vegetables can disrupt that process at its molecular root. If confirmed in clinical settings, targeting the ZNF652–GLI1 axis could open a new avenue for preventing or treating hepatocellular carcinoma, particularly in populations with high nitrosamine exposure.

Subject of Research: Sulforaphane inhibition of NDMA-induced hepatocellular carcinoma stem cell properties via the ZNF652/Sonic Hedgehog axis

Article Title: Sulforaphane inhibits N-nitrosodimethylamine induced hepatocellular carcinoma stem cell properties via ZNF652/Sonic hedgehog axis

Article References: Wu, J., Wang, X., Ren, D., Jiang, D., He, L., Chen, Y., Zhong, C., Li, X., & Han, H. (2026). Sulforaphane inhibits N-nitrosodimethylamine induced hepatocellular carcinoma stem cell properties via ZNF652/Sonic hedgehog axis. Frontiers in Cell and Developmental Biology, 14, Article 1965529. https://doi.org/10.3389/fcell.2026.1965529

Image Credits: AI Generated

DOI: 10.3389/fcell.2026.1965529

Keywords: sulforaphane, NDMA, hepatocellular carcinoma, cancer stem cells, ZNF652, Sonic Hedgehog, GLI1, broccoli, liver cancer, chemoprevention, CD133, nitrosamines

The research team, led by pathologists Beyza Erbagci and Rabia Hursitoglu together with biochemist Hasan Dagli, measured the expression of both genes in formalin-fixed, paraffin-embedded tumor tissue from 125 patients with papillary thyroid cancer. In 33 of those cases, the investigators also had access to peritumoral, non-tumoral thyroid tissue from the same patients, which served as matched control samples. Using quantitative real-time reverse transcription polymerase chain reaction, a technique that amplifies and quantifies specific messenger RNA sequences, they compared gene expression levels between tumor and healthy tissue and then searched for relationships between those levels and a battery of clinicopathological parameters, including tumor size, lymph node metastasis, lymphovascular invasion, perineural invasion, extrathyroidal extension, and tumor stage.

The first headline finding concerns what happens to the two genes inside the tumor itself. Both EGFR and GLI1 were upregulated in tumor tissue relative to the adjacent non-tumoral thyroid, confirming that these pathways are active in papillary thyroid cancer. But the more striking result was the relationship between them: rather than rising and falling together, as one might expect if the two genes were part of a single coordinated program, their expression levels showed a significant inverse correlation, with a correlation coefficient of minus 0.589 and a p-value below 0.001. In plain terms, tumors that expressed high levels of EGFR tended to express low levels of GLI1, and vice versa. That statistical signature suggests the two pathways may occupy opposing or mutually exclusive positions within the biology of these tumors, a pattern that could reflect distinct tumor subtypes or competing signaling circuits.

To understand why that inverse relationship matters, it helps to recall what each gene does. EGFR sits on the cell surface and, when activated by its ligands, triggers cascades such as RAS-MAPK and PI3K-AKT that promote cell proliferation, survival, and migration. It is a famous oncogene, amplified or mutated in many cancers, and the target of approved drugs in lung cancer and colorectal cancer. GLI1, by contrast, is a transcription factor, a protein that binds DNA and switches other genes on. It is the terminal effector of the Hedgehog pathway, an embryonic developmental program that is normally silenced in adult tissue but can be reawakened in cancer. When the pathway is active, the membrane protein Smoothened relieves inhibition of GLI proteins, allowing GLI1 to enter the nucleus and drive transcription of genes involved in cell cycle progression, stemness, and epithelial-mesenchymal transition, the process by which cancer cells acquire the mobility needed to invade and metastasize. Previous work has linked Hedgehog pathway activity to poor clinical outcomes in papillary thyroid cancer, which is precisely why the Turkish team wanted to test GLI1 as a potential marker of aggressive disease.

When the researchers cross-referenced expression levels with clinical and pathological features, the results were more restrained than the correlation finding. For most of the parameters they examined, neither EGFR nor GLI1 showed a statistically significant association. EGFR, in particular, proved disappointing as a marker: receiver operating characteristic analysis, a standard method for evaluating how well a continuous measurement discriminates between two clinical groups, showed that EGFR expression had limited performance across all the parameters tested. In other words, knowing how much EGFR messenger RNA a tumor expressed did not meaningfully help predict whether that tumor had spread to lymph nodes, invaded vessels, or extended beyond the thyroid capsule.

GLI1 told a more interesting story. While it too failed to correlate with most clinicopathological variables, it was significantly associated with one of them: microscopic extrathyroidal extension, the finding of tumor cells breaching the thyroid capsule on histological examination. This feature, reported with a p-value of 0.010, matters clinically because extrathyroidal extension is incorporated into staging systems such as the American Joint Committee on Cancer tumor-node-metastasis classification and is known to influence recurrence risk and management decisions. Beyond the single association, the ROC analysis suggested that GLI1 expression may carry discriminatory value for aggressive clinicopathological features more broadly, even though the authors characterize that ability as modest.

The authors are careful about what these results do and do not establish. Their conclusion, as stated in the paper, is that GLI1, as a key effector of the Hedgehog pathway, showed modest discriminatory ability for features associated with tumor aggressiveness in papillary thyroid cancer, and that further research is warranted to explore its therapeutic implications. That phrasing reflects a common and honest pattern in biomarker research: a gene may be biologically active and statistically interesting without being, on its own, a clinically decisive test. A single messenger RNA measurement in a heterogeneous tumor rarely captures the full complexity of pathway activation, which can depend on protein localization, ligand availability, and interactions with other signaling networks.

Nevertheless, the inverse correlation between EGFR and GLI1 is the kind of finding that can shape future hypotheses. One possibility raised by the broader literature is crosstalk between growth factor signaling and developmental pathways: in several cancer types, EGFR signaling and Hedgehog signaling have been shown to interact, sometimes cooperatively and sometimes antagonistically, depending on cellular context. If the two pathways indeed occupy alternative states in papillary thyroid cancer, tumors might be classifiable into molecular subgroups with different dependencies, and those subgroups could respond differently to targeted therapies. Hedgehog pathway inhibitors, such as smoothened antagonists already approved for basal cell carcinoma, have been explored in various malignancies, and the new data provide a rationale for asking whether a subset of thyroid tumors defined by high GLI1 expression might be candidates for such approaches.

The study also demonstrates the practical value of archival tissue. Because the investigators used formalin-fixed, paraffin-embedded samples, the standard medium of hospital pathology laboratories, their approach could be replicated in other institutions without prospective tissue collection, and the retrospective design allowed the ethics committee to waive individual informed consent given the use of anonymized archival material. The work was supported by the Scientific Research Projects Coordination Unit of Kahramanmaras Sutcu Imam University, and the authors declare no competing interests. The sample size of 125 tumors is respectable for a single-center molecular pathology study, though the 33 matched controls represent a smaller subset, and the authors’ statistical honesty about which associations reached significance and which did not is a model of careful reporting.

For patients and clinicians, the immediate takeaway is not a new test but a sharpened question. Papillary thyroid cancer is usually curable, yet the small fraction of aggressive tumors drives most of the morbidity and mortality, and the field still lacks reliable molecular tools to flag them early. This study strengthens the case that the Hedgehog pathway, acting through GLI1, participates in the biology of tumor invasion in the thyroid, while simultaneously tempering expectations about EGFR as a prognostic marker in this disease. The next steps, as the authors indicate, lie in larger and ideally prospective cohorts, in protein-level validation by immunohistochemistry, and in functional experiments that test whether manipulating GLI1 or EGFR in thyroid cancer cells changes invasive behavior. If those lines of evidence converge, the opposing expression pattern documented here could evolve from a statistical curiosity into a genuine handle on the most dangerous forms of a common cancer.

Subject of Research: EGFR and GLI1 gene expression and their association with clinicopathological features in papillary thyroid cancer

Article Title: Association of EGFR and GLI1 gene expression with clinicopathological features in papillary thyroid cancer

Article References: Erbagci, B., Hursitoglu, R., & Dagli, H. (2026). Association of EGFR and GLI1 gene expression with clinicopathological features in papillary thyroid cancer. BMC Endocrine Disorders. https://doi.org/10.1186/s12902-026-02556-9

Image Credits: AI Generated

DOI: 10.1186/s12902-026-02556-9

Keywords: papillary thyroid cancer, EGFR, GLI1, Hedgehog signaling, gene expression, biomarker, extrathyroidal extension, ROC analysis, qRT-PCR, thyroid tumor aggressiveness, oncogene, transcription factor

GLI1, short for glioma-associated oncogene family zinc finger 1, is best known as a transcription factor, a protein that switches other genes on or off. It serves as one of the main downstream effectors of the Hedgehog signaling pathway, an ancient communication system that orchestrates embryonic development and, in adults, helps maintain and repair tissues. When Hedgehog signaling is activated, GLI1 is expressed, and its presence in a cell is therefore a readout of that pathway’s activity. What has fascinated vascular biologists over the past ten years is that in adult tissues, GLI1 expression consistently marks a population of perivascular cells, the mesenchymal cells that wrap around and support blood vessels, and that these GLI1-positive cells display striking progenitor-like and injury-responsive properties.

The authors of the review, led by Weiwei Yang, Qingqing Li, Saverio Bellusci, and Xuran Chu, working across institutions in Germany and China, synthesize findings from lineage-tracing and single-cell studies in mice. Lineage tracing is a technique that permanently labels a cell and all of its descendants, allowing researchers to follow where a GLI1-positive cell goes and what it becomes over time. Single-cell technologies, meanwhile, reveal the molecular identities of individual cells within a tissue. Together, these approaches have shown that GLI1-positive perivascular cells are not a static population. They can support vessel stability and contribute to repair after injury, but under pathological conditions they can also be recruited into maladaptive remodeling programs that drive fibrosis, abnormal muscularization of small vessels, and tumor-supportive vasculature.

This duality is the central theme of the review. The key question, the authors argue, is not simply whether GLI1-positive cells are present in a tissue, but which state they have adopted and what signals push them toward one fate or another. A GLI1-positive perivascular cell that differentiates into a smooth muscle-like mural cell and stabilizes a newly formed vessel is performing a regenerative service. The same cell type, responding to chronic inflammatory or hypoxic signals, may instead proliferate excessively, deposit extracellular matrix, and contribute to the stiffened, narrowed vessels characteristic of fibrotic disease. Distinguishing these states, and identifying the molecular switches that control them, is where the therapeutic opportunity lies.

At the molecular level, the review describes how GLI1 pathway activity intersects with several major signaling networks that govern vascular behavior. The vascular endothelial growth factor, or VEGF, pathway drives endothelial cell sprouting during angiogenesis, and Hedgehog signaling has been shown to modulate the availability and effects of VEGF in several contexts. Platelet-derived growth factor signaling recruits mural cells to nascent vessels, a process in which GLI1-positive perivascular cells participate. Transforming growth factor beta signaling, a potent driver of myofibroblast differentiation and matrix deposition, represents the fibrotic arm of the same cellular program. Hypoxia-inducible factor 1 alpha, the master sensor of low oxygen, links tissue hypoxia to both angiogenic activation and, in some settings, Hedgehog pathway upregulation. The interplay among these pathways determines whether a GLI1-positive cell becomes a stabilizing mural cell, a matrix-producing fibroblast-like cell, or a contributor to pathological neovascularization.

The disease settings examined in the review span three major areas. The first is tumor angiogenesis. Solid tumors require new blood vessels to grow beyond a few millimeters in diameter, and the tumor microenvironment co-opts normal vascular biology to build them. Evidence summarized in the review indicates that Hedgehog pathway activity and GLI1-positive perivascular cells contribute to the formation of tumor vasculature in several cancer types, including glioblastoma, where glioblastoma stem cells can themselves engage Hedgehog signaling, as well as non-small cell lung cancer, triple-negative breast cancer, oral squamous cell carcinoma, and salivary adenoid cystic carcinoma. In these contexts, GLI1-positive cells may support vessel formation, stabilize tumor vessels, or contribute to the abnormal vessel architecture that impedes drug delivery while facilitating metastasis.

The second major setting is pulmonary vascular remodeling, a hallmark of pulmonary arterial hypertension and other severe lung diseases, including acute respiratory distress syndrome. In pulmonary arterial hypertension, the small pulmonary arteries undergo dramatic remodeling: smooth muscle cells proliferate, cells that should remain outside the muscle layer acquire muscle-like characteristics in a process called neomuscularization, and the vessel lumen narrows, raising pressure and eventually causing right heart failure. The review discusses evidence that GLI1-positive perivascular cells in the lung can act as progenitors for some of these aberrant smooth muscle-like cells, and that Hedgehog signaling contributes to this maladaptive program. In the injured lung, the same cells may participate in repair after acute injury, highlighting again the state-dependent nature of their behavior.

The third setting is fibrotic vasculopathy, the vascular changes that accompany organ fibrosis in tissues such as liver, kidney, skin, and lung. In the liver, hepatic stellate cells, which are perivascular mesenchymal cells by another name, drive fibrosis in chronic liver disease, and Hedgehog signaling has been implicated in their activation. The review also touches on related populations in other organs, including periodontal ligament stem cells in dental tissues and mesenchymal stem cell niches generally, where GLI1 marks cells with regenerative capacity. Across these systems, the common thread is that perivascular GLI1-positive cells sit at a decision point between repair and scarring, and the extracellular matrix, immune cells, and endothelial cells surrounding them provide the contextual signals that tip the balance.

Importantly, the authors are careful about the strength of the evidence. They explicitly distinguish between settings where direct lineage-based evidence exists, meaning that GLI1-positive cells have been formally shown to give rise to a particular cell type in vivo, and settings where the involvement of these cells is inferential, based on pathway activity or correlative data. This candor matters, because the field of perivascular cell biology has historically suffered from overlapping nomenclature and inconsistent markers. Pericytes, adventitial fibroblasts, mesenchymal stem cells, and smooth muscle cell progenitors are all perivascular mesenchymal populations with partially overlapping features, and GLI1 provides one marker among several. The review also highlights emerging tools, including spatial transcriptomics, which maps gene expression within intact tissue architecture, as a means of resolving where GLI1-positive cells reside and what states they occupy in human disease samples.

On the therapeutic front, the review identifies cell-state-specific nodes as the most realistic targets. Broad inhibition of the Hedgehog pathway with smoothened inhibitors, drugs already approved for certain cancers, has shown clinical limitations and context-dependent effects, because the pathway serves beneficial functions in tissue maintenance and repair. The more nuanced strategy proposed is to target the specific downstream programs that drive maladaptive behavior, such as the interaction between Hedgehog signaling and transforming growth factor beta-driven fibrosis, or the recruitment of GLI1-positive cells into tumor vasculature, while sparing their regenerative roles. The authors also note connections to other pathways, including YAP-dependent mechanotransduction, endothelin-1 signaling, and inflammatory nuclear factor kappa B activation, any of which could provide state-specific leverage points. As an open-access review, the article offers researchers a consolidated map of a rapidly evolving field, and it makes a compelling case that the fate of a single perivascular cell population, watched through the lens of GLI1, may determine whether an injured tissue heals or hardens, whether a tumor is fed or starved, and whether a hypertensive lung remodels relentlessly or can be coaxed back toward stability.

Subject of Research: The role of GLI1-positive perivascular cells in angiogenesis and vascular remodeling

Article Title: GLI1⁺ perivascular cells in angiogenesis and vascular remodeling: mechanisms and disease relevance

Article References: Yang, W., Li, Q., Liu, X., Yang, W., Noori, A., Sotoodeh, L., Chen, M., Xiong, X., Bellusci, S., & Chu, X. (2026). GLI1⁺ perivascular cells in angiogenesis and vascular remodeling: mechanisms and disease relevance. Cellular and Molecular Life Sciences. https://doi.org/10.1007/s00018-026-06412-w

Image Credits: AI Generated

DOI: 10.1007/s00018-026-06412-w

Keywords: GLI1, Hedgehog signaling, perivascular cells, angiogenesis, vascular remodeling, tumor angiogenesis, pulmonary arterial hypertension, fibrosis, lineage tracing, pericytes, VEGF, TGF-beta