Glioblastoma (GBM), a grade IV treatment refractory aggressive brain tumor, is one of the deadliest forms of cancer with an average patient survival of 15 months and 5-year survival rate of less than 5%. This poor prognosis is despite an aggressive standard-of-care consisting of maximally safe surgical resection followed by radiation and temozolomide treatment. Standard-of-care has not changed over the past two decades. There is a clear unmet need for better therapeutic strategies to combat GBM.
Genomic profiling of GBMs identified genetic alterations within three pathways: receptor tyrosine kinase (RTK)/Ras/PI3K; p53/ARF/MDM2/MDM4; and RB/CDK4/INK4A. Hyperactivation of the RTK/Ras/PI3K pathway leads to the persistent activation of diverse serine/threonine signaling networks that promote proliferation, metabolic reprogramming, and resistance to standard-of-care treatment. In addition, GBMs possess an enhanced signaling heterogeneity and plasticity which facilitates the activation of downstream STK signaling pathways. The underlying mechanisms that enable this signaling adaptability in GBMs remain poorly understood.
A significant knowledge gap in our understanding of aberrant oncogene signaling is the role of serine/threonine phosphatases, in particular protein phosphatase 2A (PP2A). PP2A is a bona fide tumor suppressor, physiologically functioning as the counterbalance to the serine/threonine signaling networks activated downstream of RTK/Ras/PI3K activation. Unlike other tumor suppressors, such as p53 (~35% genetic alteration) and PTEN (~36% genetic alteration), PP2A (<1% genetic alteration) is genetically intact in GBM. However, it remains unclear how GBMs evade the tumor suppressor function of PP2A to maintain persistent activation of serine/threonine signaling networks. In this work, our goal was to elucidate regulatory mechanism(s) exploited by GBM to overcome the tumor suppressor activity of PP2A, with the aim of identifying actionable therapeutic targets. We took two independent approaches to investigate the mechanisms of PP2A suppression in GBM. First, we evaluated a group of endogenous inhibitors of PP2A (EIP) with abnormally high expression in GBMs. Second, we investigated drug repurposing strategies using FDA-approved antipsychotics with PP2A activating and anti-GBM properties.
Aim1: Endogenous inhibitors of PP2A drive tumorigenesis, maintain the activation of oncogenic serine/threonine signaling hubs, and promote radiation resistance by preventing PP2A targeting DNA damage response kinases.
We identify three EIPs, acidic leucine-rich nuclear phosphoprotein 32 family member A (ANP32A), cancerous inhibitor of PP2A (CIP2A), and nuclear oncogene SET (SET), with increased expression in GBMs compared to normal brain tissue. To investigate the potential oncogenic function of these EIPs, we utilized CRISPR-Cas9 gene editing to generate knockouts in established and patient-derived GBM cell lines. EIP knockout effectively enhanced PP2A activity. Intracranial xenograft models bearing tumors from ANP32A, CIP2A, and SET knockout cells had improved survival compared to the non-target control cells. As a result, we investigated the role of these EIPs in disrupting the tumor suppressor function of PP2A in GBM.
We set out to determine the phospho-protein networks protected by ANP32A-, CIP2A-, and SET-dependent inhibition of PP2A in GBM. PP2A is a heterotrimeric enzyme comprising three subunits, structural, regulatory, and catalytic, each subunit possessing several isoforms. An impediment to studying the PP2A interactome is the complexity of PP2A composition (> 80 trimer complexes), posttranslational modifications such as phosphorylation (phosphatase inhibition)/methylation (phosphatase activation), and instability outside of in situ conditions. To overcome these limitations, we engineered a proximity biotin ligase system by fusing the second-generation biotin ligase (BioID2) to the PP2A structural subunit. The BioID2 enzyme facilitates the biotinylation of proteins in direct proximity to PP2A and enables the tracking of the PP2A interactome under in situ conditions. Subsequently, we used the BioID system to measure the PP2A interactome in EIP knockout cells using unbiased proteomics. This revealed that GBMs rely on EIPs to maintain the activation of specific oncogenic signaling nodes with important functions in survival, proliferation, and treatment-resistance. SET is required to maintain the activating phosphorylation of Akt, PKA C, and STAT1 as well as the inhibitory phosphorylation of GSK3. ANP32A maintains the activating phosphorylation of transcription factors NFkB p65 and LYRIC. CIP2A is necessary to protect cMyc from degradation and maintain the activating phosphorylation of STAT1 and STAT3. Together, we conclude that EIP-dependent inhibition of PP2A is a critical determinant of serine/threonine signaling hub activation in GBM.
In addition, we discover that ANP32A, CIP2A, and SET promote radiation resistance by preventing PP2A inhibition of the DNA damage response (DDR) kinases ATM, ATR, and CHK1. The proteomic analysis of the PP2A interactome identified an increase in PP2A-ATR and/or PP2A-ATM interaction in ANP32A, CIP2A, and SET knockout cells. These interactions were maintained following radiation, resulting in the suppression of ATM and ATR activating phosphorylation. We confirmed these increased interactions in unirradiated and irradiated cells using in situ proximity ligation assays (PLA). In the EIP knockout cells, suppression of radiation-induced activation of ATM and/or ATR impaired single-strand and double-strand DNA damage repair. In addition, we observed a repression of radiation-induced activation of CHK1, ablating the G2/M checkpoint. Together, decreased DNA damage repair and G2/M checkpoint control caused radiation-induced mitotic cell death. As a proof-of-concept, we demonstrate that ATM and CHK1 inhibitors radiosensitize control cells without enhancing the radiosensitization of EIP knockout cells. This indicated that ATM and CHK1 suppression are key contributors to the radiosensitizing effect of EIP loss. We conclude that the aberrant expression of EIPs in GBM prevents PP2A targeting of DDR kinases, preventing radiation induced mitotic cell death by facilitating DNA damage repair and proper G2/M cell cycle checkpoint control.
Together, we illustrate the necessity of PP2A inhibition by three endogenous inhibitory proteins, ANP32A, CIP2A, and SET, for GBMs to maintain persistent activation of serine/threonine kinases and transcription factors with critical functions in growth factor and DNA damage response signaling pathways. Our results highlight the therapeutic potential of PP2A activation; capable of direct and simultaneous suppression of multiple oncogenic signaling hubs.
Aim2: FDA-approved antipsychotics obstruct vesicle homeostasis, inducing lysosomal cell death in GBM.
Our second approach focused on the repurposing of FDA-approved compounds for central nervous system (CNS) disorders, with PP2A activating properties. Our investigation focused on a group of schizophrenic drugs called phenothiazines reported to directly activate PP2A. We demonstrate that phenothiazines (perphenazine, fluphenazine, chlorpromazine, and acetophenazine) have anti-GBM properties across a panel of commercially available and patient-derived cell lines. Among the first- and second-generation phenothiazines tested, perphenazine possessed the most potent anti-GBM properties. Daily perphenazine treatment significantly decreased intracranial xenograft growth and increased the survival of nude mice. In turn, we investigated the anti-GBM mechanism of action of perphenazine.
In coordination with a large body of published work, we demonstrate that perphenazine increases PP2A activity in GBM. However, antagonizing the activation of PP2A using small molecule inhibitor LB-100, viral oncoprotein (SV40 ST), CRISPR Cas9 deletion and gene silencing of specific PP2A subunits did not rescue cellular viability following perphenazine treatment. Attempts to use a recently developed phenothiazine analog DT-061, reported to activate specific PP2A subunits, was also unsuccessful. We redirected our efforts to understand the PP2A independent anti-GBM property of perphenazine.
We report that in GBM, perphenazine induces cellular vacuolization, a phenotypic change consistent with the disruption of vesicle trafficking pathways. This corresponded with a blockade of autophagic and endocytic flux, sensitizing GBM cells to diverse metabolic stress. We determine that the disruption of vesicle trafficking is caused by perphenazine induced neutralization and inactivation of lysosomal function. Using immuno-affinity pulldown of lysosomes, we demonstrate that perphenazine accumulates in lysosomal vesicles using mass spectrometry. The accumulation of perphenazine in the lysosome leads to lysosomal membrane permeabilization (LMP) and cell death. Pre-treatment of cells with the pan v-ATPase inhibitor bafilomycin A1 (BafA1), which blocks the electrochemical gradient required for drug transport into the lysosome, ablated perphenazine accumulation in the lysosome, inhibited LMP, and rescued cell death. Therefore, the anti-GBM activity of perphenazine is related to the perturbation of lysosomal homeostasis.
Perphenazine is highly lipophilic and a weak base; compounds with similar properties are reported to accumulate within lysosome vesicles. Therefore, we postulated that the anti-GBM activity of drugs approved for CNS disorders may be predicted by their chemical properties. As a proof-of-concept, we demonstrate that the anti-GBM effect of commonly prescribed 2nd- and 3rd-generation antipsychotics and antidepressants is predicted by lipophilicity, pKa, and steric hinderance of the most basic nitrogen. Effective compounds include aripiprazole (Abilify™), sertraline (Zoloft™), fluoxetine (Prozac™), duloxetine (Cymbalta™/Yentreve™), cyclobenzaprine (Flexeril™/Amrix™), and amitriptyline (Elavil™). Similar to perphenazine, these drugs decreased lysosomal activity and induced LMP in GBM cells. Pre-treatment with BafA1 rescued LMP and cellular viability of all effective 2nd- and 3rd-generation antipsychotics and antidepressants. Therefore, the anti-GBM activity of antipsychotics and antidepressants relates to their ability to disrupt lysosomal homeostasis, inducing lysosomal cell death. We conclude that antipsychotics and antidepressants with lysosome disrupting properties may be good candidates for drug repurposing to combat GBM.