Background: How do benign nerve tumors become malignant?

A neural crest cell model derived from human induced pluripotent stem cells (iPSCs) reproduced early features of malignant peripheral nerve sheath tumor (MPNST) development. The researchers sequentially inactivated genes to analyze changes in cell identity and confirmed that olaparib plus selumetinib suppressed tumor growth in mice. The research was limited to cell and animal experiments.

In neurofibromatosis type 1 (NF1), benign plexiform neurofibromas (PNFs) can progress to malignant tumors. Loss of function in NF1, CDKN2A, and polycomb repressive complex 2 (PRC2) is involved in this process, but models for studying the successive effects of these changes on cell differentiation and malignant transformation have been limited. The researchers modeled early progression by editing CDKN2A in NF1-deficient iPSCs and then inactivating SUZ12, a PRC2 component.

Key findings: Loss of glial differentiation capacity and selective drug responses

Cells with NF1 deficiency and loss of function of both p14ARF and p16INK4a, encoded by CDKN2A (double knockout, 2KO), were differentiated toward the Schwann cell lineage. They were then cocultured as spherical aggregates with neurofibroma-derived fibroblasts and implanted into the sciatic nerves of immunodeficient mice. Neurofibroma-like tumors were observed four months later, but histologic atypia was absent. The model therefore cannot be considered to fully reproduce the stage of atypical neurofibromatous neoplasm of uncertain biologic potential (ANNUBP).

Cells with SUZ12 also inactivated (triple knockout, 3KO) acquired mesenchymal stem cell-like features. Analyses of chromatin accessibility and gene expression showed reduced promoter accessibility and expression of SOX10, a regulator of glial differentiation, alongside activation of neurogenic and mesenchymal programs. Differentiation experiments also showed that the cells had lost the ability to become Schwann cells. Implantation of 3KO cell spheroids into the sciatic nerve produced tumors consistent with early MPNST, though they were less aggressive than patient-derived orthotopic xenograft tumors.

The researchers treated three-dimensional neural crest spheroids with 339 small molecules that modulate epigenetic regulation and assessed viability after 48 hours. Poly(ADP-ribose) polymerase (PARP) inhibitors showed selective activity in PRC2-deficient spheroids. The combination of the PARP inhibitor olaparib and the MEK inhibitor selumetinib also showed synergy in 3KO spheroids, but its effect was limited in two-dimensional cultures of established MPNST cell lines. Because disease stage and culture conditions both differed between the comparison models, the difference in response cannot be explained by PRC2 status alone.

Significance and outlook: A combination candidate evaluated in a single animal model

In an exploratory experiment using the patient-derived orthotopic xenograft (PDOX) model NF1-18B, in which patient-derived tumor tissue was implanted into the sciatic nerve, groups of 5–8 mice were treated for 21 days. Olaparib alone did not clearly change tumor volume or growth patterns compared with the control group. By contrast, relative tumor volume in the group receiving the combination with selumetinib was 61.6% lower than in controls, a statistically significant difference. This figure does not mean that tumors shrank by 61.6% from their size before treatment began.

No weight loss or damage to the examined liver, kidney, lung, or heart tissues was observed during the observation period. However, there is no clinical evidence for assessing patient safety, and differences in response according to mouse sex remain a subject for further investigation. Since the results come from a cell model reproducing only some aspects of early malignant transformation and from a tumor derived from a single patient, further animal experiments and optimization of dosing conditions are the next steps.

The methods and results are presented in the full Nature Communications paper.