Targeting Glioblastoma Through Phospholipase D
By: Carol A. Rouzer, VICB Communications
Published: December 2, 2013
Inhibition of phospholipase D in glioblastoma cells impedes Akt signaling, leading to failure of autophagy and cell death.
Glioblastoma multiforme (GBM, Figure 1) is the most common brain tumor in adults, comprising approximately 15% of all brain tumors. Fast growing and aggressive, GBM is nearly always fatal, with a median survival time of 14 months after diagnosis. Current treatment for GBM, including a combination of surgery, radiation therapy, and adjuvant chemotherapy, is at best, only minimally effective. Attempts to discover more efficacious, targeted drugs for GBM have been hindered by a high degree of genetic variability in the tumor, rapid acquisition of drug resistance, and the inability of molecules to penetrate the blood-brain barrier. Undaunted by these obstacles, Vanderbilt Institute of Chemical Biology members Alex Brown and Craig Lindsley, along with their laboratories, have identified phospholipase D (PLD) as a promising new target in the battle against GBM [R. C. Bruntz, et al. (2013) J. Biol. Chem., published online November 20, doi:10.1074/jbc.M113.532978].
Figure 1. Side view of the brain of a patient with glioblastoma multiforme as seen by magnetic resonance imaging. Image reproduced with permission from Wikimedia Commons, under the Creative Commons Attribution-Share Alike 30 Unported license.
The two PLD isoforms, PLD1 and PLD2, catalyze the hydrolysis of phosphatidylcholine (PC) to phosphatidic acid (PtdOH). PLD1’s basal activity is low, and the enzyme requires interactions with ADP-ribosylation factor (ARF), Rho family members, protein kinase C, and other signaling components to be activated. In contrast, PLD2 is constitutively active and postulated to be regulated in cells by release from inhibition. PLDs play an important role in numerous signaling pathways, and elevated PLD activity is found in many forms of cancer, where it is associated with increased anchorage-independent growth, invasiveness, and tumorigenesis. Yet, despite PLD’s wide-ranging functions in cell signaling and lipid metabolism, genetic deletion of either PLD isoform in mice produces no obvious deleterious phenotype. These considerations led the Brown and Lindsley labs to hypothesize that PLD inhibition could be an effective therapeutic approach for some forms of cancer.
Observations that transfer of U87MG cells (a GBM cell line) to serum-free conditions elicited an increase in PLD activity suggested that PLD might play a role in the response to environmental stress in GBM. This conclusion was supported by the finding that treatment of serum-free cultures with VU0359595, a PLD1 isoform-preferring inhibitor, or VU0365739, a PLD2 isoform-preferring inhibitor, resulted in both reduced PLD activity and decreased cell viability. The finding that VU0365739 inhibited PLD activity and cell viability with greater potency than VU0359595 suggested that PLD2 was the most important isoform in the response to serum-free culture, a conclusion that was supported by siRNA silencing of each isoform. Glioma stem cell clones derived from patients also showed reduced viability in the presence of PLD inhibitors when grown under high stress conditions. Even when cultured in complete medium, the PLD inhibitors reduced anchorage-independent growth of these cells, suggesting that a reliance on PLD signaling is a common feature of GBM cells.
The phosphatidylinositol 3-kinase (PI3K) signaling pathway is activated in a high percentage of GBMs. PI3K converts phosphatidylinositol-4,5-bisphosphate (PIP2) to phosphatidylinositol-3,4,5-trisphosphate (PIP3). PIP3 serves as a membrane anchoring point for a number of signaling proteins, which bind to PIP3 through their plextrin homology (PH) domain. An important effector of PI3K signaling is the Akt protein, which binds to PIP3 along with its activating kinase, phosphoinositide-dependent kinase 1 (PDK1). PDK1 phosphorylates Akt at threonine-308. Other kinases, including mTORC2 (mammalian target of rapamycin complex 2) phosphorylate Akt at serine-473. The phosphorylated Akt then plays a role in multiple signaling pathways that regulate cell growth, metabolism, proliferation, and survival. Cessation of PI3K signaling occurs through the function of PTEN (phosphatase and tensin homolog), which dephosphorylates PIP3 to yield PIP2 (Figure 2).
Figure 2. Activation of the PI3K/Akt pathway. In this example, a receptor tyrosine kinase (RTK) binds its ligand (L) leading to activation of the receptor. Receptor activation leads to autophosphorylation at key tyrosine residues. PI3K binds to the RTK at one of these phosphotyrosines, resulting in its activation in proximity of the membrane. It now can convert its substrate PIP2 (magenta hexagons) to PIP3 (green octagons). Akt binds to PIP3 in the membrane through its plextrin homology domain. PDK1 also binds to PIP3, and phosphorylates Akt at threonine-308. Other kinases, including mTORC2, phosphorylate Akt at serine-473. Phosphorylated Akt is now activated. PTEN converts PIP3 back to PIP2 as the first step to reverse the signaling process.
In cancer, activation of PI3K signaling can occur at many stages of the pathway. The U87MB cells used by the Brown and Lindsley labs lacked an active PTEN enzyme, eliminating the primary off switch for the PI3K/Akt pathway. This led the researchers to explore the possible interactions between PLD and Akt in U87MB cells. They discovered that PLD inhibition by small molecule inhibitors or siRNA knockdown, under conditions of serum starvation, substantially decreased the levels of activated Akt. Immunoprecipitation experiments demonstrated that PLD2 directly interacts with Akt, but this interaction was not affected by the PLD inhibitors. By contrast, addition of PtdOH to PLD inhibitor-treated cells reversed the effects of the inhibitors on Akt activation and cell viability. The investigators concluded that it was the product of the PLD-catalyzed reaction that was critical to cell survival under serum-free culture conditions.
Binding studies revealed that PtdOH binds directly to Akt and increases the kinase’s affinity for PIP3. PtdOH binding required the PH domain of Akt, but a mutation that eliminates the PIP3-PH domain interaction had no effect on PtdOH binding. These results indicated that Akt binds to both lipid species at separate sites on its PH domain, and suggested that PtdOH should facilitate Akt’s interaction with PIP3 in the membrane. Consistently, PLD inhibitors decreased the amount of total Akt and phosphorylated Akt in U87MG cell membranes, an effect that was reversed by addition of PtdOH.
PLD inhibitor treatment of serum-starved U87MG cells did not trigger death by apoptosis, leading the investigators to search for an alternative mechanism to explain the PLD inhibitor-mediated loss of cell viability. Autophagy is a process by which cells isolate damaged intracellular components into autophagosomes. These vesicles fuse with lysosomes, leading to digestion of the contents and nutrient recycling. Failure of normal autophagic processing is associated with cell death under some circumstances. Indeed, monitoring of two markers of autophagy revealed that PLD inhibition resulted in failure of autophagosome processing, while autophagosome formation was unaffected.
To explain the effects of PLD inhibition on autophagy, the investigators focused on Beclin1, a protein phosphorylated by Akt. Beclin1 and Rubicon (RUN-domain cysteine-rich domain-containing, Beclin1-interacting protein) form a complex that inhibits autophagosome processing. Phosphorylation of Beclin1 by Akt prevents its interaction with Rubicon, thereby relieving the complex-mediated suppression of autophagic flux. Consequently, PLD inhibition, leading to failure of Akt activation, should result in increased Beclin1/Rubicon complex formation and reduced autophagosome processing (Figure 3). Consistently, the investigators found that PLD inhibitors increased the amount of Beclin1 bound to Rubicon in U87MG cells.
Figure 3. Proposed mechanism for the effect of PLD inhibitors on GBM cell survival. (TOP) Activation of Akt occurs as described in Figure 2, except that generation of PtdOH (blue starbursts) from PC (red spheres) by PLD2 facilitates binding of Akt to PIP3 in the membrane. This promotes activation of Akt, which then phosphorylates Beclin1. Phosphorylated Beclin1 is unable to bind to Rubicon, and in the absence of this complex formation, autophagic flux can proceed. (Bottom) In the presence of a PLD inhibitor (PLDi), no PtdOH is generated. As a result, Akt does not bind to PIP3 in the membrane and cannot be activated. Beclin1 remains unphosphorylated, so it binds to Rubicon, and the complex inhibits autophagic flux. Failure of autophagy leads to cell death.
U87MG cells expressing a constitutively active form of Akt were resistant to PLD inhibitor-mediated cell death. In these cells, PLD inhibitors did not block Akt activation or increase Beclin1/Rubicon complex formation, and their effect on autophagic processing was markedly reduced as compared to processing in U87MG cells expressing wild-type Akt.
Together, the results indicate that, at least under stress conditions, GBM cells rely on PLD2 to increase Akt activation as a protective mechanism. An important outcome is Akt-dependent phosphorylation of Beclin1, leading to stress-related autophagic flux. Failure of PLD2-mediated Akt activation leads to a loss of cell viability. The finding that the nontumorigenic HEK293 cell line did not exhibit PLD-dependent Akt activation under serum-free conditions suggests that these observations are associated with the malignant phenotype. The results also suggest that PLD2 may be a valuable target for blocking the growth and survival of GBM in vivo.