These findings link the mechanistic logic of AP-1 network wiring to a well-known adaptive resistance phenotype in melanoma and show how model-guided perturbations can block this transition.

Under MAPK inhibition, the network adapts; loss of JUND, down-regulation of FRA1, and up-regulation of cJUN, promoting a FRA2-High state that drives dedifferentiation (loss of MITF and SOX10) and MAPK inhibitor resistance.

The model, trained on data from 19 genetically-distinct melanoma populations, reveals quantitative rules explaining why some clones occupy few AP-1 states while others span broad, heterogeneous configurations.

Now, we move from description to mechanism!
We combined high-throughput multiplexed single-cell protein measurements with a mechanistic ODE model of dimerization-controlled, co-regulated, and competitive AP-1 interactions.

In our previous study (Comandante-Lou et al., Cell Reports 2022), we introduced the idea of an “AP-1 state”, the relative balance of AP-1 factors predicting melanoma cell differentiation and plasticity at single-cell resolution. pubmed.ncbi.nlm.nih.gov/35926467/

Each AP-1 member competes for limited binding partners, engages in auto- and cross-regulatory feedback, and exhibits signal-dependent stability and transcriptional control.

Among regulators of plasticity, AP-1 transcription factors (FOS and JUN families) stand out. They form a densely interconnected network of dimers integrating diverse MAPK signals.

Cell state plasticity, the ability of genetically-identical cells to adopt distinct, switchable phenotypes, underlies both normal processes like differentiation and cancer traits like metastasis, immune evasion, and therapy resistance.