More Precious Than Gold? Platinum Inhibitors Target Aβ

A new approach to detoxifying the amyloid-β (Aβ) peptide borrows from an unlikely source—the field of cancer drugs. In the May 7 PNAS online, Kevin Barnham, Roberto Cappai, and colleagues from the University of Melbourne, Australia, report on a new class of platinum-based inhibitors of Aβ aggregation and toxicity.

The inhibitors are designed to target the high-affinity metal binding site of Aβ, where the presence of copper or zinc can contribute to Aβ’s damaging effects. While questions will need to be addressed about the specificity and safety of these particular compounds, the results present a novel approach to interfering with Aβ via blocking the residues responsible for metal binding.

For some time, the Australian researchers have been interested in inhibiting the interaction between Aβ and metals as a therapeutic strategy. That work resulted in the clinical testing of the metal chelator clioquinol in AD patients. After promising early results (see ARF related live discussion and ARF related news story), clioquinol was discontinued, but the second-generation drug PBT2 is still in the running (see ARF related drug story).

In the new work, Barnham and coworkers take a different tack. Rather than trying to tie up the metal, they target Aβ itself with the goal of blocking its metal-binding site with platinum. To do that, the investigators replaced two small amine groups that complex the platinum in the widely used cancer drug cisplatin with a polyaromatic phenanthroline structure. This scaffold is related to other Aβ binders including PIB, Congo red and thioflavin T, and contacts aromatic residues in the N-terminal region of Aβ42 near the metal-coordinating histidine side chains.

Structural studies using three versions of phenanthroline-platinum complex indicate that the new compounds bind to Aβ in vitro and alter its structure. Mass spectrometry and NMR pointed to platinum coordinating with the metal-binding histidine residues 6, 13, and 14. Circular dichroism measures suggested that the compounds prevented Aβ from adopting a β-sheet structure in the presence of copper.

The compounds affected Aβ toxicity as well. First, the platinum complexes inhibited two in-vitro correlates of Aβ toxicity, the time-dependent aggregation of Aβ42 and copper-mediated H2O2 production. The compounds were as potent at blocking peroxide production as clioquinol. Next, the researchers showed that the compounds prevented Aβ toxicity in mouse primary neuronal cultures, and one of the compounds was also shown to reverse the synaptic effects of Aβ on long-term potentiation in mouse hippocampal slices.

“The results from this work have given us the confidence that if we can design compounds that can get to the target in vivo, then we can reasonably expect that they are capable of inhibiting the deleterious effects of Aβ,” Barnham wrote to ARF in an e-mail. However, he continued, “The compounds described in the paper are unlikely to have the necessary properties to achieve this, so we have undertaken a chemistry program to design molecules capable of being orally administered and inhibiting Aβ in vivo. This work is ongoing and we have been able to synthesize Pt-based novel chemical entities and have begun studies in transgenic animals. Our initial studies have given us cause for optimism as we have seen changes in Alzheimer disease-related biomarkers such as Aβ oligomers and tau phosphorylation.”

The new compounds, Barnham wrote, have been patented and assigned to Prana Biotechnology Ltd. of Parkville, Victoria, Australia, the company that tested clioquinol and is developing PBT2. Barnham, Cappai, and their coauthors Ashley Bush, Colin Masters, and Robert Cherny are all disclosed on the PNAS paper as consultants to Prana.—Pat McCaffrey

Comments

  1. This is an interesting approach. The study appears to be well conducted overall, but I would respectfully note that the SELDI mass spectrometry result could require further validation. If one just wants to measure the mass of Aβ by MALDI-TOF mass spectrometer, then the SELDI platform used here is appropriate. But the authors used MALDI-TOF mass spectrometry to measure the stochiometry and binding of platinum to Aβ (Figure 1). My experience would suggest that this is problematic because the complex is no longer in the solution phase and is instead measured in the solid and gas phases. MALDI-TOF mass spectrometry would not typically be the method of choice to study non-covalent ligand-protein interactions. Most mass spectrometrists would prefer to use electrospray ionization (ESI-MS) to study ligand-protein interactions, because this method will measure the mass of complex from the solution phase into the gas phase.

    In addition, I am a bit concerned that the Aβ/Pt complex was treated in strong acid (0.5 percent in TFA) during the SELDI measurement, as described in Page 5 of the manuscript. This strong acid and low pH treatment is rather non-physiological and not an ideal condition one would normally use to measure the protein complex by mass spec. A further look at various different mass spec instrumentation platforms and ionization methods could advance this research, and also explain why there appears to be a discrepancy between mass spec and NMR results.

    View all comments by Austin Yang

  2. We'd like to respond to Austin Yang’s comment. The purpose of the NMR and mass spectroscopy data presented in Fig. 1 was not so much to measure the stoichiometry of the Aβ/drug adduct as to identify that adducts were formed, and to confirm that the binding site on Aβ was indeed the histidine residues as was predicted.

    Pt has some unique properties that make its potential use targeting biological molecules quite attractive. One of these is that Pt is considered a kinetically inert metal (i.e., substitution reactions are very slow). One consequence of this is that the adducts Pt forms are very stable even to treatment by acidic conditions and the harsh conditions of the mass spectrometer. For all practical purposes, the Pt-Aβ bond can be regarded as a covalent bond.

    View all comments by Kevin Barnham

  3. This is novel and interesting work; however, it raises questions concerning approved AD targets and recent progress in the understanding of DNA repair.

    Clioquinol, which was advanced for AD therapy by some of the authors, failed for metal chelation-associated side effects. Now, a compound that utilizes a related mode of action, albeit with subtle yet significant differences in complex geometry, is based on more stable, square planar platinum II complexes. This looks promising at first glance. Square planar complexes versus equilibria of tetrahedral or square planar Cu(I)/Cu(II) complexes have a much higher chance of interfering with Aβ aggregation, as most aggregation inhibitors are fully planar.

    The caveats I see at this stage derive from both the ligand and the platinum core. A bisulfonated ligand is unlikely to pass the blood-brain barrier. Both platinum and the phenanthroline come with the liability of gene toxicity, as a literature of 362 phenanthroline PtCl2 complexes in 223 publications substantiates. Regarding the specific probes used in this publication: the core phenanthroline PtCl2 complex is known to intercalate with DNA. It forms nucleotide dimers and is likely to inhibit error-prone polymerase eta or related polymerases (Damsma et al. 2007). Incidentally, reference 6 used in this manuscript (Cleare M et al., Bioinorganic Chemistry (1973), 2(3), 187-210) is quite outdated now. The phenyl substitution makes the compound even more cytotoxic. For these reasons, I'd consider these poor leads for an AD therapy (De Pascali et al., 2006; Moeller et al., 2000).

    The luminescence of such Pt complexes consumes thiols (Zuleta et al. Journal of the American Chemical Society (1989), 111(24), 8916-17). In follow-up work, it might be helpful to use controls to rule out such luminescence-derived activity. Furthermore, it would be nice to see controls for the thioflavin fluorescence of the target in the presence of Pt-complexes, which is dramatically different, of course.

    Importantly, comet or micronucleus assays to address genotoxicity would be helpful to assess whether there is a safe therapeutic window for carcinogenic or cytotoxic compounds. I would respectfully submit that the paper's conclusion based on the inactivity of cis-platin seems premature to me, as this compound is not luminescent. However, the conclusion related to the histidine complexation as a therapeutic approach is sound. On balance, these compounds may be useful for imaging or as tools, but I doubt they qualify as leads for AD therapy in their present form.

    References:

    Damsma GE, Alt A, Brueckner F, Carell T, Cramer P. Mechanism of transcriptional stalling at cisplatin-damaged DNA. Nat Struct Mol Biol. 2007 Dec;14(12):1127-33. PubMed.

    De Pascali SA, Migoni D, Papadia P, Muscella A, Marsigliante S, Ciccarese A, Fanizzi FP. New water-soluble platinum(II) phenanthroline complexes tested as cisplatin analogues: First-time comparison of cytotoxic activity between analogous four- and five-coordinate species. Dalton Trans. 2006 Nov 14;(42):5077-87. PubMed.

    Moeller N, Kangarloo BS, Puscasu I, Mock C, Krebs B, Wolff JE. Rational design of platinum chemotherapeutic drugs: hydrophobicity increases cytotoxicity. Anticancer Res. 2000 Nov-Dec;20(6B):4435-9. PubMed.

    View all comments by Boris Schmidt

  4. Having read this interesting paper, it seems to me that at this point, these types of molecules are not druggable. Looking at the structures, I would ask whether these compounds are toxic and insoluble. Follow-up work could specifically address potential toxicities of such a framework or template that is polyaromatic, planar, lipophilic—parameters that could raise potential concerns of a chemical series.
    This is important as there is literature precedent alluding to various toxicities of transition metals and 1,10-phenanthroline systems. It is advantageous to address potential toxicity liabilities, in this case chromosomal aberrations or carcinogenicity, of the chemical leads early on in the discovery process. Finally, the authors note that solubility is an issue with compound 2 and 4; I wonder how one could try to improve that parameter for compound 3.

    View all comments by Corinne Augelli-Szafran

  5. Both Corrine Augelli-Szafran and Boris Schmidt raise concern with toxicity for the class of molecules examined in our paper, and these concerns are entirely valid. That said, they are valid for any new drug paradigm whatever molecular scaffold is being investigated. I have spent 10 years working with Pt anti-cancer drugs. We are well aware of the baggage that comes with this class of molecule, and close attention will be paid to issues such as genotoxicity, etc.

    I'd like to challenge some inaccuracies in Boris Schmidt’s comment. First, clioquinol did not fail as an AD therapy because of “metal chelation-associated side effects.” That this misconception persists is disappointing—it is a pharmacological equivalent to an urban myth. As has been explained previously on Alzforum, manufacturing glitches led Prana Biotechnology, the company funding the trial, to halt it and throw its resources behind developing the second generation of similar compounds. That strategy has been successful, as the lead compound PBT2 has successfully completed a Phase 2a trial (see ARF related drug story). No adverse events were reported in the trial.

    Second, the chemical and biological properties of the Pt-based anti-cancer drugs such as cisplatin are different from those of the Pt(phenanthroline) compounds. The Pt(phenanthroline) compounds are not effective anti-cancer compounds. I find the comment with respect to the toxicity of the Pt(phenanthroline) compounds to be misleading. It implies that the Damsma et al. 2007 manuscript describes these compounds as such. Yet the only compound the paper describes is cisplatin; there is no mention of any phenanthroline derivatives. The statement: “The phenyl substitution makes the compound even more cytotoxic” is interesting as it is followed by a reference to De Pacali et al., 2006. This reference actually shows that the phenanthroline compounds are two to three orders of magnitude less toxic than cisplatin. The toxicity observed for the 1,10phenathroline compounds was observed at concentrations of 200-500 micromolar, hardly physiologically relevant concentrations.

    Like with most of the assays, the structurally different cisplatin was used as a control for the inhibition of amyloid formation as reported by ThT fluorescence. Unlike the Pt(phenanthroline) compounds, cisplatin did not inhibit amyloid formation. All the ThT amyloid inhibition results were confirmed by electron microscopy. These data were reported in the paper, although not shown in the figures.

    View all comments by Kevin Barnham

Make a Comment

To make a comment you must login or register.

References

Therapeutics Citations

  1. Clioquinol

Webinar Citations

  1. Clioquinol Trial Postmortem: Does Blocking Metal-Aβ Interactions Work?

News Citations

  1. Pilot Study Suggests Clioquinol Benefits AD Patients
  2. Anti-Amyloid Drug Clears Phase 2a Hurdle

Other Citations

  1. PBT2

Further Reading

Primary Papers

  1. Barnham KJ, Kenche VB, Ciccotosto GD, Smith DP, Tew DJ, Liu X, Perez K, Cranston GA, Johanssen TJ, Volitakis I, Bush AI, Masters CL, White AR, Smith JP, Cherny RA, Cappai R. Platinum-based inhibitors of amyloid-beta as therapeutic agents for Alzheimer's disease. Proc Natl Acad Sci U S A. 2008 May 13;105(19):6813-8. PubMed.

Annotate

To make an annotation you must Login or Register.