p53-MDM2 Inhibition

“D-peptide inhibitors of the p53–MDM2 interaction for targeted molecular therapy of malignant neoplasms”

Overview

Figure 1. Structure of nutlin-3 molecule previously discovered as an MDM2 antagonist.

Many studies have focused on finding MDM2 antagonists such as nutlin-3 (Figure 1), which has shown p53-dependent destruction of cancer cells at high concentrations. Mirror-image phage display has revealed the discovery of p53-MDM2 interaction inhibitors, DPMI-α (TNWANLEKLLR) and mutant DPMI-β (TAWYANFEKLLR), that compete with p53 for MDM2 with an even higher binding binding affinity than nutlin-3 [28]. Further selectivity has resulted in the discovery of DPMI-γ (DWWPLAFEALLR) which is more potent than the previous two [27]. Being D-peptides, all are resistant to proteolysis, however they fail to cross the cell membrane alone to induce p53 dependent tumor cell death. Efforts have been made to use liposomes as carriers, sacs of phospholipids known for carrying drugs and toxins across the membrane. These MDM2 and MDMX antagonists provide hope for therapeutic treatments of certain cancers.

Mirror Image Phage Display

A synthetic biotin-MDM2 protein complex composed completely of D-amino acids was screened with a peptide phage library.  Those that interacted were amplified in E. Coli (ER2738). 7 of 10 binding clones contained the sequence for DPMI-γ (Figure 2), verifying the sequence as competitively binding to MDM2 [27].

Figure 2. Results from mirror image phage display after four rounds of selection showing ten sequences that effectively interacted at the MDM2-p53 binding site. A conensus of seven of ten sequences (bolded) matched that of D-PMI-γ and therefore verify the D-protein competitive antagonist for MDM2.

Structural Validation of DPMI-γ as p-53 activator

Using the Phaser program, the researchers developed a cocrystal structure of MDM2 and DPMI-γ to further understand the interaction on a structural level. DPMI-γ forms a left handed helix that interacts with MDM2 primarily through hydrophobic interactions in the binding pocket: DTrp2, DTrp3, DPro4, DPhe7, DLeu10, and DLeu11 (Fig 3). Further, while the MDM2 for both PMI, the L-enantiomer, and DPMI-γ is nearly identical, the amino acid sequence of DPMI-γ only shares 25% of its amino acid sequence with PMI [27]. For example, an important hydrophobic triad of Phe3/Trp7/Leu10 in PMI is actually DTrp3/DPhe7/DLeu11 in DPMI-γ. Even then, the DPMI sequences are still both comparable to those in p53, helping them interact with MDM2 at the same site as p53 and show that both are competitive inhibitors. DPMI-γ is also distinct from DPMI-α, however, these differences are not very energetically different and the binding affinities of both molecules are nearly identical. The advantage of course for the D-peptides over PMI is not only that the D-peptides have higher binding affinity, but that they are resistant to proteolysis and thus a higher bioavailability would be ideal for cancer treatment.

Figure 3. Crystal structure of MDM2 binding site containing p53-MDM2 interaction inhibitor, DPMI-γ. (A) Surface model of MDM2 containing DPMI-γ. Hydrophobic residues that interact with MDM2 are highlighted. (B) Ribbon drawing of important hydrophobic residues of DPMI-γ that interact with MDM2 binding pocket. Oxygen molecules are in red whereas Nitrogen molecules are in blue. PDB ID: 3IWY

DPMI Toxicity and Dependence on p53

Although resistant to proteolysis with a much higher binding affinity than nutlin-3, DPMI-γ was unable to kill colon cancer cells by itself likely because it cannot cross the cell membrane.  When conjugated with nine D-Arginine residues at its C terminus, the peptide rapidly killed the cancer cells, however this was determined to be through factors that were p53 independent. A detergent-like molecule was likely created that was toxic to all cell types, which has been proven in previous studies [27]. Liposomes were therefore used as carriers for D-peptides to cross the cell membrane. DPMI-α was used for this study and was also bound to a cellular recognition molecule, RGD, to target tumor cells on the cell surface. Cells with wild type p53 and RGD-bound DPMI-α showed higher levels of dose-dependent tumor growth inhibition than those without the RGD (Figure 4). Mutant p53 cells also showed DPMI-α-RGD inhibition, proving that DPMI-α was p53 dependent in function [27]. Western blots further supported this finding.

Figure 4.  Growth inhibition of cancer cells by nutlin-3, free DPMI-α, liposome-DPMI-α, RGD-liposome-DPMI-α, and RGD-liposome-LPMI-α using the MTT cell viability assay. Three measurements were averaged to acquire inhibition curves. Reproduced with permission from Liu, M.; Li, C.; Pazgier, M.; Li, C.; Mao, Y.; Lv, Y.; Gu, B.; Wei, G.; Yuan, W.; Zhan, C.; Lu, W.Y.; Lu, W. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 14321–14326.

When tested on cancer-containing mice over a two week period, DPMI-α-RGD treatments showed drastic improvements in tumor growth inhibition that helped the animals live longer. The most important discovery was that a high concentration of D-peptide was not needed to be successful, and the molecule also had a higher binding affinity than the other well known inhibitor, nutlin-3.

Summary

The authors develop several high-potency D-peptide inhibitors of the p53-MDM2 interaction. While many studies have focused on nutlin-3 as a viable p53-MDM2 inhibitor, the discovery of D-PMI’s have provided molecules with much higher efficiency in binding. These molecules still have room for improvement in selectivity and MDM2 binding affinity, but their resistance to proteolysis and high bioavailability have made them strong candidates as therapeutic agents.