AKT Inhibition Modulates H3K4 Demethylase Levels in PTEN-Null Prostate Cancer

Abstract

Hyperactivated AKT kinase due to loss of its negative regulator PTEN influences many aspects of cancer biology, including chromatin. AKT primarily regulates acetyl-CoA production and phosphorylates many histone-modulating enzymes, resulting in their activation or inhibition. Therefore, understanding the therapeutic impact of AKT inhibition on chromatin-related events is essential. Here, we report that AKT inhibition in prostate-specific PTEN knockout mice significantly induces di- and trimethylation of H3K4 with concomitant reduction in H3K9 acetylation. Mechanistically, we observed that AKT inhibition reduces expression of the H3K4 methylation-specific histone demethylases KDM5 family, especially KDM5B expression at transcriptional levels. Furthermore, we observed that AKT negatively regulates miR-137 levels, which transcriptionally represses KDM5B expression. Overexpression of miR-137 significantly reduced KDM5B and increased H3K4 methylation levels but failed to change AKT phosphorylation. Overall, we observed that AKT transcriptionally regulates KDM5B mainly via repression of miR-137. Our data identify a mechanism by which AKT kinase modulates the prostate cancer epigenome through regulating H3K4 methylation. Additional studies on AKT inhibition-mediated induction of H3K4 methylation will help in designing strategies to enhance the therapeutic efficacy of PI3K/AKT inhibitors.

Figure 1.

AKT kinase inhibition modulates global epigenome marks and induces H3K4 methylation. A, Histograms show fold change in H3 histone modifications after treatment with AKT inhibitor perifosine using the Epiquick Histone Modification Multiplex Assay kit in PTEN deficient in vivo samples. Error bars illustrates SEM B, Western blot analysis (n=3) of tissue extracts isolated from wild type and PTEN-knock out mice samples after perifosine treatment showing expression pattern of H3K4Me1/2/3 proteins C, Western blot analysis (n=3) of perifosine-treated PTEN null PC3 cell lines at 6 hr and 24 hr time point. Total H3 histone was used as loading control. D, Representative photomicrographs (magnification, 10 & 40X) showing immunohistochemical staining for p-AKT and caspase 3 in Pten wild type and knockout tumor sections of vehicle- or perifosine-treated mice.

Figure 2.

AKT kinase inhibition reduces histone demethylase KDM5B. A, qPCR analysis of selective HDMs in AKT inhibitor treated in vivo samples showing mRNA expression patterns.18S actin was used as loading control B, Western blot analysis (n=3) of tissue extracts isolated from wild type and PTEN-knock out mice samples in control as well as perifosine-treated conditions in duplicate. Vinculin was used as a loading control. Bars represent normalized data (mean±SEM, n=3), *P<0.05 C, PC3 cells transfected with non-target siRNA and siRNA targeting AKT1 as well as AKT2. In another set of experiment PC3 cells were treated with perifosine for 3 hr, 6 hr and 24 hr. qPCR analysis of both experimental set ups showed expression pattern of KDM5B mRNA. 18S actin was used as a loading control D, Western blot analysis of whole cell lysates isolated from PC3 cell after treatment with perifosine for 6 hr and 24 hr. Vinculin was used as a loading control. Histograms represent quantification of western blot results normalized to loading control (mean±SEM, n=3), *P<0.05 E, in vivo histone demethylase (KDM5) activity assay of the nuclear lysates after perifosine treatment.

Figure 3.

AKT negatively regulates miR-137 expression to transcriptionally regulate KDM5B levels. A, Histograms represent pattern of miR-137 expression in samples isolated from perifosine treated mice against control samples showed by qPCR. 18S actin was used as loading control B, Histograms showing RNA levels of KDM family from PC3 cells transfected with miR-137 over expression constructs and miR-137 inhibitor by qPCR C, PC3 cells were transfected with miR-137 over-expressing constructs and whole cell lysates were analysed by western blotting to study the status of KDM5B protein along with p-AKT and H3K4 methylation marks. Vinculin was used as loading control. Bars depicting normalized data (mean±SEM, n=3), *P<0.05.

Figure 4.

Combination of AKT inhibitor with broad methyltransferase inhibitor synergize to reduce PTEN null prostate cancer tumorigenicity. Effect of combination of AKT inhibitor (perifosine) and DZNeP (broad histone methylation inhibitor) on PTEN null PC3 tumor growth in athymic nude mice. Approximately 1 million PC3 cells were subcutaneously injected in each flank of the mouse to initiate tumor growth. Once tumors started to grow, their sizes were measured twice weekly and the tumor volume was calculated A, Average tumor volume of vehicle, perifosine, DZNeP and combination treated mice plotted over weeks after tumor cell inoculation. Values represent mean ± SE of six animals. **p<0.001 versus the vehicle treated group of mice. Details are described in Materials and Methods B, Representative photographs of tumors bearing mice from each group.