Melanoma’s Hidden Architect Builds Blood Vessels and Shuts Down Immune Defences

Before you were born, a protein called HOXD13 helped shape your fingers and toes. It did its job during embryonic development, switched off, and — in most people — was never heard from again. But in roughly half of all skin melanomas, something goes wrong. HOXD13 wakes back up. And it turns out the same molecular programme that once built your limbs can also build a tumour’s blood supply while simultaneously shutting down the immune cells trying to destroy it.

That’s the finding from a team led by Eva Hernando-Monge and Pietro Berico at NYU Langone Health, published in Cancer Discovery on 30 January. They’ve identified HOXD13 as a kind of master coordinator, a transcription factor that doesn’t just flip one genetic switch but reaches across the genome, pulling distant regulatory elements into three-dimensional loops that activate whole networks of genes at once. The result is a tumour that can feed itself and hide from the immune system through a single reactivated developmental programme.

“Our study provides new evidence that transcription factor HOXD13 is a potent driver of melanoma growth and that it suppresses the T cell activity needed to fight the disease,” said Berico, a postdoctoral fellow at NYU Grossman School of Medicine and the Perlmutter Cancer Center. What’s striking about the finding, published in Cancer Discovery on 30 January, is the sheer breadth of what HOXD13 controls. It doesn’t just flip one switch. It reaches across the genome, binding to distant regulatory elements — stretches of DNA far from the genes themselves — and pulling them together in three-dimensional loops that activate whole networks of genes at once.

The team analysed tumours from more than 200 melanoma patients across the US, Brazil and Mexico, screening more than a thousand transcription factors to see which ones stood out in melanoma compared with benign moles and other cancer types. HOXD13 was the only one strongly linked to both increased blood vessel formation and reduced immune cell infiltration. It wasn’t restricted to any particular melanoma subtype either; the protein appeared in tumours regardless of their genetic driver or cellular identity state.

The researchers traced HOXD13’s reactivation back to melanoblasts — the embryonic precursor cells that eventually become melanocytes, the pigment-producing cells in our skin. In those early developmental cells, HOXD13 is normally active. It shuts off once melanocytes mature. But when an oncogenic mutation like BRAF V600E strikes, HOXD13 can switch back on in neural crest cells and melanoblasts — though not in mature melanocytes — sort of rebooting a foetal programme in cells that should have left it behind. This reactivation appears to precede the formation of distinct tumour cell states, which suggests it is an early event in melanoma’s progression rather than a consequence of it.

So what exactly does HOXD13 do once it’s back in business? Three downstream targets proved particularly important. The first two — VEGFA and SEMA3A — are both involved in angiogenesis, the sprouting of new blood vessels. VEGFA promotes vessel growth, while SEMA3A modulates how those vessels mature. Together, they appear to create an abnormal vasculature: leaky, disorganised blood vessels that can’t function properly as conduits for immune cells. When the researchers silenced HOXD13 in mouse tumours, they saw the vascular architecture shift — more endothelial cells, fewer pericytes (the support cells that wrap around blood vessels), and less of the fibrin leakage that signals dysfunctional vessels.

The third target is perhaps the cleverest bit. HOXD13 also activates a gene called NT5E, which encodes a protein called CD73. This enzyme sits on the surface of melanoma cells and chops up extracellular molecules to produce adenosine — a chemical that acts as an immunological stop sign. Adenosine puts the brakes on T cells, preventing them from entering the tumour and killing cancer cells. It’s a molecular shield, essentially. Blood levels of cytotoxic T cells were lower in melanoma patients with high HOXD13 activity than in patients without the cancer or without overactive HOXD13, and the ability of those T cells to infiltrate tumours was reduced too.

The most dramatic experiments involved forcing HOXD13 expression in mouse melanoma cells that don’t normally produce much of it. In mice with functioning immune systems, the tumours exploded in size. In immunocompromised mice, the growth advantage vanished — strong evidence that HOXD13’s main benefit to the tumour comes from its ability to hide from immune attack, not from making cancer cells divide faster. Fluorescent imaging of those tumours told a stark visual story: control tumours were dotted with CD4 and CD8 T cells; HOXD13-overexpressing tumours were virtually barren of them.

But if HOXD13 is orchestrating this two-pronged assault through known downstream pathways, that creates a therapeutic opening. “This data supports the combined targeting of angiogenesis and adenosine-receptor pathways as a promising new treatment approach for HOXD13-driven melanoma,” said Hernando-Monge, a professor in the Department of Pathology at NYU Grossman School of Medicine. Her team treated HOXD13-overexpressing tumours in mice with a combination of lenvatinib (which blocks VEGF receptors) and etrumadenant (which blocks adenosine receptors). Neither drug alone did much. Together, they dramatically reduced tumour growth, shrank the abnormal vasculature, and — crucially — restored immune cell infiltration, with T cells and macrophages flooding back into the tumour.

Separate clinical trials are already underway evaluating VEGF-receptor and adenosine-receptor inhibitors for melanoma and other cancers. If those prove successful, Hernando-Monge says her team plans to initiate clinical investigation of the combination approach, specifically targeting melanoma patients whose tumours show elevated HOXD13. She also wants to explore whether VEGF and adenosine pathways could be targets in other HOXD13-positive cancers, including some glioblastomas and sarcomas.

There are caveats, of course. HOXD13 turns up in roughly half of melanoma patients, so this wouldn’t be a universal strategy. And the mouse models could only run for a limited time before tumours ulcerated, leaving questions about long-term durability. Still, the work opens an intriguing new angle on one of cancer’s oldest tricks — how a tumour can build itself a blood supply while simultaneously making sure the immune system never gets an invitation to visit.

Study link: