Melanoma: How does the skin environment influence its progression?

Melanoma is a skin cancer that develops from melanocytes, the cells that produce melanin. Although it is not the most common form of skin cancer, it is the leading cause of skin cancer-related mortality due to its high metastatic potential. Its incidence continues to increase in connection with exposure to UV radiation and the ageing of the population. Despite major advances in medicine over recent decades, melanoma remains difficult to control, notably due to the development of treatment resistance.

These phenomena are not explained solely by the emergence of new mutations, but also rest on substantial cellular plasticity, allowing tumour cells to change phenotype in response to their environment.

Tumour cells are indeed in constant interaction with their surroundings, comprising stromal cells, immune cells and the extracellular matrix made up of collagen, fibronectin and other structural proteins. This matrix does not merely play a passive supporting role but actively influences cell behaviour, proliferation, migration and survival. In melanoma, the extracellular matrix undergoes profound remodelling during tumour progression and in response to treatments. An increase in the amount of collagen and a rearrangement of its fibres lead to a well-documented phenomenon of tumour tissue stiffening.

This increased stiffness is now recognised as a major biophysical characteristic of aggressive tumours.

It is within this framework that the recent work of DECKERT and his team on melanoma is set. Their approach rests on a simple but long-overlooked hypothesis: the rigidity of tumour tissue may not affect all melanoma cells in the same way. More precisely, a cancer cell’s ability to sense and interpret the mechanical constraints of its environment may depend on its phenotypic state. To test this idea, the researchers chose to experimentally recreate environments of controlled rigidity by culturing different populations of melanoma cells on collagen matrices of varying compliance. This approach yielded a first key finding: when confronted with extracellular matrix stiffness, not all melanoma cells respond in an equivalent manner.

To test the hypothesis that extracellular matrix stiffness does not affect all melanoma cells in the same way, DECKERT and his team studied human melanoma cell lines exhibiting either a melanocytic phenotype or a dedifferentiated phenotype. To isolate the effect of matrix stiffness, the cells were cultured on highly compliant collagen matrices (below 1 kPa), comparable to soft tissues, or, conversely, on very rigid matrices (above 16 kPa), mimicking fibrotic tumour tissue.

The initial morphological observations revealed a striking difference between the two cellular states. When cultured on a rigid matrix, dedifferentiated cells exhibit a pronounced cell spreading and an increased adhesive surface area, indicative of active interaction with their environment. Conversely, melanocytic cells maintain a relatively consistent morphology, whether grown on a soft or rigid matrix.

This difference in behaviour was confirmed during functional analysis of cell proliferation. Using real-time monitoring, the researchers demonstrated that matrix stiffness strongly stimulates the proliferation of dedifferentiated cells. Conversely, a compliant matrix induces cell cycle arrest in these cells, associated with a decrease in the expression of key cell cycle progression proteins, such as phosphorylated retinoblastoma protein (P-Rb) or the transcription factor E2F1. This effect is much more modest, or even absent, in differentiated melanocytic cells.

To evaluate their invasive potential, cells were cultured for several days on soft or stiff matrices and then tested in Boyden chamber invasion assays coated with a standard extracellular matrix (Matrigel). The results show that prior exposure to a stiff matrix markedly enhances the ability of de-differentiated cells to invade the Matrigel, whereas it has little effect on melanocytic cells. These data suggest that matrix stiffness does not act only transiently, but induces a genuine pro-invasive functional programme in de-differentiated cells.

Another important point: these differences in response are observed irrespective of the cells’ mutational status.

Whether they harbour mutations in the BRAF gene or the MEK gene, which are common in melanoma cells, dedifferentiated cells consistently remain more sensitive to mechanical cues than melanocytic cells, demonstrating that the response to extracellular matrix stiffness is dictated not by genetic alterations but by the phenotypic state of the cells.

Finally, the researchers investigated the consequences of this mechanosensitivity on the response to targeted therapies. By culturing cells on soft or stiff matrices and exposing them to a combination of BRAF and MEK inhibitors (Dabrafenib and Trametinib), they demonstrated that rigidity confers marked protection to dedifferentiated cells, which become less susceptible to apoptosis induction. Conversely, a soft matrix resensitises these cells to treatment, with a significant increase in caspase 3 activation and cell death. Once again, this effect is specific to dedifferentiated cells and is not observed in melanocytic cells.

The researchers concluded that the ability of melanoma cells to exploit the rigidity of their environment is closely dependent on their phenotypic state.

Dedifferentiation not only confers invasive traits and resistance to therapies, it also involves a dependence on mechanical signals from the extracellular matrix. One may then ask the following question: by what molecular mechanisms do these cells perceive collagen stiffness and translate this physical information into biological signals that favour their aggressiveness? It is precisely this point that scientists have sought to elucidate in their subsequent studies.

Cells detect the mechanical properties of their environment through transmembrane receptors that interact with components of the extracellular matrix. While integrins are the most studied matrix receptors in this context, attention here has focused on another family of collagen receptors: the discoidin domain receptors DDR1 and DDR2. These receptors are unique in that they are activated by fibrillar collagen and are involved in transmitting mechanical signals within cells.

The researchers first analysed the expression and activation of these receptors across the different melanoma cell states. They noted that, while several collagen receptors are expressed in dedifferentiated cells, only DDR1 and DDR2 show a significant increase in activation when the cells are grown on a rigid matrix. This activation is much weaker, or even absent, in differentiated melanocytic cells, which suggests that DDR1 and DDR2 could play a role in the mechanosensitivity of dedifferentiated cells.

To answer this question, the researchers blocked DDR1 and DDR2 in two different ways: by reducing their expression or by inhibiting their activity. In both cases, they observed that the cells lose their ability to respond to stiffness. They generate fewer internal forces, their cytoskeleton becomes disorganised and their matrix-anchoring structures are less developed. These results show that DDR1 and DDR2 are essential for activating the cell’s contractile machinery, on which the activation of mechanosensitive signalling pathways depends, particularly those involving the transcriptional co-activator YAP. When the cells are subjected to high mechanical stresses, YAP accumulates in the nucleus and activates the expression of genes involved in proliferation, migration and cell survival.

Researchers therefore examined the transcriptional activity of YAP in melanoma cells cultured on substrates of differing rigidity. The results show that, in dedifferentiated cells, a stiff substrate induces a marked nuclear translocation of YAP, accompanied by an increase in the expression of its target genes. Conversely, on a soft substrate, YAP remains predominantly cytoplasmic and its transcriptional activity is greatly reduced. This response is far less pronounced in melanocytic cells, confirming that YAP activation by rigidity is closely linked to the cells’ phenotypic state. Moreover, when DDR1 and DDR2 are inhibited, this activation of YAP almost completely disappears, even in the presence of a stiff substrate.

These studies thus enable the reconstruction of the following transmission chain: collagen stiffness is sensed by DDR1 and DDR2, relayed by cellular contractility, and then transduced by YAP activation.

However, one question remained unanswered: in melanoma cells, how does YAP remain active when it is normally rapidly degraded? This led researchers to investigate another layer of regulation: the stability of proteins and the mechanisms of ubiquitination.

Under normal conditions, YAP activation is transient: after fulfilling its role as a regulator of gene expression, this protein is rapidly targeted for degradation by the ubiquitin–proteasome system. This mechanism serves to temporally limit the strength of mechanosensitive signals. However, within a rigid extracellular matrix, as observed in melanoma, YAP remains persistently active. To understand this abnormal persistence, researchers opted to move beyond classical signalling pathways and focus on the mechanisms that control protein stability. They thus explored the role of deubiquitinases (DUBs), enzymes capable of removing ubiquitin chains attached to proteins and shielding them from proteasomal degradation.

In practice, dedifferentiated melanoma cells were cultured on soft or rigid collagen matrices, then lysed to analyse their enzymatic activity. The cell extracts were incubated with a modified ubiquitin probe designed to bind specifically to the active site of functional deubiquitinases. The labelled enzymes were then isolated and identified. This approach revealed that the activity of several DUBs depends on the stiffness of the extracellular matrix. Among them, the enzyme USP9X stands out clearly: its activity increases significantly when cells are exposed to a rigid matrix, whereas it decreases in a softer environment.

Researchers then sought to understand what triggers the activation of the USP9X enzyme in a stiff environment. To do so, they blocked cytoskeletal contractility, that is, the cell’s ability to generate internal forces via the actin–myosin complex. When they inhibit myosin II, USP9X is no longer activated. This demonstrates that USP9X activation depends directly on the mechanical tension generated within the cell. Similarly, when DDR1 and DDR2 receptors are blocked, USP9X fails to activate, even on a rigid matrix.

Further tests conducted by scientists have shown that USP9X acts directly on YAP. When USP9X activity is blocked, YAP is rapidly tagged with ubiquitin and degraded by the proteasome, even in the presence of a rigid matrix. Conversely, when USP9X is active, it removes these degradation marks, allowing YAP to accumulate within the cell.

The stiffness of the tissue therefore does more than simply activate YAP: it prevents its degradation. USP9X stabilises YAP and prolongs the mechanical response of dedifferentiated cells. This stabilisation explains why these cells remain migratory, invasive and resistant to treatments for as long as their environment remains rigid.

To verify that this mechanism is not limited to observations in vitro, the researchers then assessed the role of USP9X in an animal model of melanoma. They used bioluminescent melanoma cells to track their behaviour in real time within the organism. These cells, either controls or rendered USP9X-deficient by genetic inhibition, were injected intravenously into immunodeficient mice. This model enables the study of the very earliest stages of metastatic dissemination, particularly the ability of tumour cells to exit the bloodstream and colonise the lungs.

Just hours after injection, bioluminescence imaging revealed a clear difference between the two groups. Cells lacking USP9X exhibited a markedly reduced ability to extravasate from the blood vessels and to implant in lung tissue. Longitudinal follow-up of the animals over nearly two months confirmed this initial observation: whereas mice injected with control cells gradually developed pulmonary metastases, no detectable metastases were observed in animals receiving USP9X-deficient cells.

These results demonstrate that USP9X is essential for the early stages of migration and invasion required for metastasis formation, consistent with its role in stabilising YAP.

The team of scientists then sought to determine whether targeting USP9X could also interfere with the mechanical reprogramming induced by targeted therapies. It is known that inhibition of the BRAF–MEK pathway, while initially effective, promotes remodelling of the extracellular matrix, increased tumour stiffness and sustained activation of YAP, contributing to relapse. To this end, BRAF-mutant murine melanoma cells were injected into immunocompetent mice, which were then treated either with targeted therapy alone, with the USP9X inhibitor alone (referred to as G9 in the study), or with the combination of both.

As expected, the USP9X inhibitor administered alone slightly slows tumour growth without inducing marked regression. In contrast, when USP9X inhibition is combined with targeted therapy, tumour relapse is significantly delayed and animal survival is improved.

Tumour analysis reveals that this combination prevents the fibrotic remodelling typically induced by BRAF and MEK inhibitors. Collagen networks are less dense, deubiquitinase activity is reduced, and activation of YAP, along with that of its target genes involved in tumour cell migration, invasion and resistance, is markedly decreased. These findings demonstrate that blocking USP9X interrupts the self-sustaining mechanical loop established by the treatments, preventing YAP stabilisation and the mechanical adaptation of tumour cells.

These studies primarily offer a new perspective on our understanding of melanoma.

They show that tumour cell adaptation does not rely solely on genetic or transcriptional mechanisms, but also on their ability to sense their environment and to stabilise certain proteins such as YAP over time. By identifying USP9X as an intermediary between extracellular matrix rigidity, mechanosignalling and therapeutic resistance, these findings pave the way for new strategies targeting the mechanisms that enable melanoma cells to adapt and persist in the body.

These findings also suggest that targeting the mechanical response of melanoma could complement existing therapeutic approaches. In particular, USP9X inhibition represents an indirect means of limiting sustained YAP activation and of curbing the mechanically-driven reprogramming induced by targeted therapies. More broadly, this strategy could apply to other cancers developing in rigidified tissues, where YAP plays a central role, such as lung cancer.

That said, these studies present some limitations highlighted by the researchers themselves. The experiments are primarily based on simplified collagen matrices, which do not replicate the full complexity of the extracellular matrices produced in vivo by tumour and stromal cells. Moreover, the precise mechanisms linking tissue stiffness, collagen receptor activation and the increase in USP9X activity remain only partially elucidated. These studies therefore represent a first step, calling for further research to explore these mechanisms in more physiological environments and to determine the extent to which they can be exploited clinically.

• LAVERSANNE M. & al. Global burden of cutaneous melanoma in 2020 and projections to 2040. JAMA Dermatology (2022).

• OLIVIA A. G. & al. Skin-on-a-chip technology: Microengineering physiologically relevant in vitro skin models. Pharmaceutics (2022).

• DECKERT M. & al. USP9X is a mechanosensitive deubiquitinase that controls tumor cell invasiveness and drug response through YAP stabilization. Cell Reports (2025).

• DECKERT M. & al. Extracellular matrix stiffness determines the phenotypic behavior of dedifferentiated melanoma cells through a DDR1/2-dependent YAP mechanotransduction pathway. Research Square (2025).