Cellular identity crisis: Antiandrogen resistance by lineage plasticity

ABSTRACT

A recent publication in Science demonstrates the ability of prostate cancer cells to switch lineages from one that is dependent on androgen signaling to a cell type that is not. Known as lineage plasticity, this phenomenon is driven by the transcription factor SOX2 in RB1 and TP53-deficient prostate cancer. SOX2 is a potential prognostic marker and therapeutic target in castration resistant prostate cancer.

Prostate cancer persists as one of the leading causes of cancer-related deaths among men in the United States. Antiandrogen therapy is initially effective in treating castration resistant prostate cancer (CRPC), but resistance ultimately occurs. In their recent publication in Science,¹ Mu et al. demonstrate a novel mechanism of antiandrogen resistance, termed “lineage plasticity,” in which prostate tumor cells undergo dedifferentiation to take on luminal, basal, and epithelial characteristics and display a neuroendocrine (NE) phenotype. The appearance of NE cells in prostate cancer is known to be associated with aggressive tumors and poor prognosis.² While luminal cells depend on the androgen receptor (AR) for growth, NE cells often develop during antiandrogen therapy to bypass AR-targeted therapies. However, the mechanism that underlies lineage plasticity in antiandrogen-treated prostate cancer is poorly understood.

Large-scale genomic analysis of alterations found in clinical CRPC revealed that mutations in 2 tumor suppressors, TP53 and RB1, were common in clinical NE tumors (74%) and in cases that developed resistance to enzalutamide or abiraterone (84%).¹ Mu et al. confirmed that reducing both TP53 and RB1 expression via shRNA targeting resulted in the greatest level of in vitro and in vivo enzalutamide-resistance. Downstream AR effectors continued to be inhibited by enzalutamide despite continued cellular proliferation, suggesting that the new cell phenotype was no longer dependent on the AR for growth.

Mu et al. next demonstrated TP53 and RB1 loss leads to lineage plasticity in LNCaP/AR cells, which is characterized by high expression of basal and NE markers and lower expression of luminal genes. Using a doxycycline inducible model, they showed that basal and NE marker expression increased with the introduction of shRB1/shTP53 and decreased once shRNA was no longer present. When RB1 and TP53 expression was reduced, lineage plasticity permitted prostate cancer cells to grow in the presence of enzalutamide. Therefore, TP53 and RB1 loss is sufficient to confer lineage plasticity and enzalutamide resistance, and lineage plasticity resulting from RB1 and TP53 loss is reversible.

Continued large-scale genomic analysis showed higher expression of a transcription factor, SOX2, in CRPC with TP53 and RB1 alterations. SOX2 promotes pluripotency during development and is a marker of poor prognosis and metastasis in several different cancers.³ Mu et al. confirmed that SOX2 expression increased when TP53 and RB1 expression was reduced in in vitro and in vivo models. A reduction of SOX2 expression in addition to RB1 and TP53 reversed the expression of NE markers and restored enzalutamide sensitivity. These results demonstrate that SOX2 drives lineage plasticity and antiandrogen resistance when TP53 and RB1 are dysfunctional.

Tumors with intact TP53 and RB1 are more likely to develop antiandrogen resistance through restoration of AR signaling, while tumors with altered TP53 and RB1 are more likely to develop resistance via lineage plasticity. Unchecked SOX2 reprograms cells to enter anAR-independent progenitor-like state that circumvents antiandrogen therapy. Wildtype TP53 has been found to inhibit SOX2 through expression of targeted micro RNAs, while wildtype RB1 inhibits SOX2 by creating repressive chromatin marks in the SOX2 promoter.^4,5

The use of large-scale genomic data to examine cohorts of patients exhibiting more aggressive disease and/or treatment resistance is a powerful method to discover novel genetic variations with predictive and prognostic importance. Using this method, Mu et al. uncovered a highly important insight into the subset of CRPC demonstrating TP53 and RB1 loss: cells lacking these tumor suppressors exhibit lineage plasticity as a function of the SOX2 pathway to circumvent antiandrogen therapy. Since these alterations are found in approximately 39% of CRPC adenocarcinoma, 74% of neuroendocrine CRPC, and 84% of CRPC previously treated with antiandrogens, the present data are likely to have a significant clinical influence if validated in human patients.

First, these findings add to a growing body of evidence suggesting that SOX2 expression is associated with poor prognosis and metastasis in several cancers.³ The results of Mu et al. demonstrate retrospective evidence that such tumors are more frequent in more severe disease; however, it remains unclear whether patients with these mutations have a worse prognosis than those not carrying TP53 and RB1 loss. Second, Mu et al. provide convincing evidence that SOX2 expression secondary to TP53 and RB1 loss is associated with antiandrogen resistance in vitro and in vivo. In ovarian cancer, SOX2 overexpression is similarly associated with treatment resistance to paclitaxel.⁶ However, future clinical studies are required to demonstrate that these findings could guide the selection of a variety of therapies available in the pharmaceutical armamentarium. The predictive and prognostic significance of AR-V7 and TMPRSS-ERG are comparable translational interests. Patients with AR-V7-expressing CRPC survived longer on taxanes instead of antiandrogens, and TMPRSS-ERG is associated with poor prognoses.⁷

In addition to its value as a marker, SOX2 may also be a viable drug target. Vaccination with SOX2 peptides in mice with oligodendroglioma has demonstrated a delay in tumor development and lethality.⁸ The results of Mu et al. suggest that future studies should determine if certain therapies that directly target SOX2 are effective as well. SOX2 is involved in a complex network of signaling cascades, including VEGF, EGFR, PI3K/Akt, Myc, Hedgehog, mTOR and Wnt. Several pharmacological inhibitors of these signaling cascades exist, and future study could determine if these inhibitors influence SOX2 expression. Since RB1 inhibits SOX2 via an epigenetic mechanism, it is possible that DNA methylation inhibitors, histone deacetylase inhibitors, and histone methyltransferase inhibitors are possible therapeutic options. A recent study showed that the inhibition of a DNA methylase, Ezh2, restored antiandrogen sensitivity in prostate cancer cells exhibiting lineage plasticity.⁹ Ezh2 inhibitors are already in various phases of development for several cancer types. These potential routes of therapy give hope for the treatment of CRPC patients demonstrating lineage plasticity.

Disclosure of potential conflicts of interest

The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products or organization imply endorsement by the US Government.

Funding

This work was supported by the National Institute of Health (NIH) under grant 1ZIABC010547–14.

Agency: National Institute of Health (NIH)

type: Investigator-Initiated Intramural Research Projects (ZIA)

Project #: 1ZIABC010547–14

Application #: 9343646