A friend in the business called my attention to this paper, which is about another piece of the ubiquitination system that I was writing about here just the other day – in this case, the deubiquitinating enzyme Rpn11. There are a couple of classes of deubiquitinators – some of them use a cysteine in their active site, and some of them have a metal atom. Rpn11 is a metalloprotease, and it’s involved in the handling of proteins that have been ubiquitinated and are at the gate of the proteasome, the cell’s shredding machinery for proteins that have been so tagged. It strips off a ubiquitin as the marked proteins move further into the proteasome, which is a tidy way of handling things, since that can be re-used while the rest gets broken down for recycling at the amino acid level (ubiquitinated proteins are, in general, too wide to even make it through the space available unless this happens). Loss-of-function mutations in this enzyme are lethal; the proteasome mechanism depends on this step. If too many ubiquitinated proteins start piling up uncleared, it’s generally a signal for apoptosis, the cell death program, to kick in.
However, gumming up the proteasome in a less comprehensive manner is a known strategy for cancer therapy, thanks to Velcade (bortezomib) and the drugs that have followed it. That one’s reknowned as the first boronic-acid-based protease inhibitor to actually make it as a drug, after years of misses (or not even bothering) with that particular functional group. It actually binds to an active site in the complex, further down from that Rpn11-mediated gate, and shuts down the proteasome completely. This, as it turns out, is especially lethal to myeloma tumor cells, and Rpn11 inhibitors might well do something similar (and do it even to cells that have mutated their way into becoming resistant to bortezomib, since it’s a different mechanism of action.
So far, so good. There are indeed such things as metalloprotease inhibitors, plenty of them, and they generally work by tying up that key metal ion. And here’s where the trouble starts. This new paper is all about the role of Rpn11 in myeloma cell lines, and one key to that argument is showing that inhibition of the enzyme leads to trouble for them. The authors went back to a 2002 paper and used their reported inhibitor for these current studies – a molecule called OPA. It does indeed cause apoptosis in myeloma cell lines, even ones that are resistant to bortezomib.
But here’s what makes me a bit nervous. “Inhibitor” is not some sort of catch-all class, a switch that gets thrown or something. OPA is known to chemists as ortho-phenanthroline, and it’s one of the most common and dependable metal-complexing molecules known. Those two nitrogen atoms are fixed in just the right “bite” for a lot of different metal ions; you find this property being used all the time in metal-organic and inorganic chemistry. In that 2002 paper, it was used on purified proteasomes that had been removed from cells, and was presented as evidence that the protease involved in a particular subunit (Rpn11, as we know it) was indeed a metalloenzyme. I have no problem with that – toss a universal metal chelator into an in vitro system, and you’ll disturb anything that depends on metal ions.
This current paper, though, takes the “here’s an Rpn11 inhibitor” idea and runs with it. The compound is introduced, with a reference to the earlier paper, just as “a metallopeptidase inhibitor. . .that inhibits Rpn11 activity”. It’s used in cell assays (several different kinds), and is also dosed systemically (at 20 mg/kg ip) in mice with tumor xenografts. That, to me, is pushing it a bit. o-Phenanthroline surely does a lot of things in cells, and in whole animals, too (I should note that its effects on myeloid cell lines as a potential chemotherapy agent were first reported back in 1985). To be sure, this current paper did show that it had much less cytotoxic activity against normal mononuclear cells, but I think that referring to it as an “Rpn11 inhibitor”, as the paper does for its entirety, is going too far. I’d be happier if it were identified as a nonselective metal chelator, because that’s what it is – a very crude tool indeed, and one that is surely causing other things to happen in living systems that may or may not confound these results. The natural killer activity of lymphocytes was reported many years ago to be affected by phenanthroline, probably through zinc coordination.
Phenanthroline itself does not show up in the Chemical Probes Portal, which has recently been relaunched, but it and its derivatives are certainly all over the literature. There are many papers where it’s added to cell assays or to rodent experiments as something to coordinate zinc or iron, in a nonspecific manner. You see it used a lot as well when people are trying to develop metal complexes as therapeutic agents, but a control experiment (which doesn’t always get run) is seeing whether just the metal-free ligand is having the effect. Here’s a group, for example, that was studying various phenanthroline-based complexes of ruthenium as anticancer agents – they found that the ligands themselves, minus the metal, were more active than the metal complexes they’d started with. (For that matter, here’s a very recent report of such metal phenanthroline complexes being active against multiple myeloma cell lines – you wonder if they need the metal, too, or if, in light of this paper under discussion, it’s just hurting what should be a dose of plain old phenanthroline!)
Overall, I think that this new paper is probably right. Rpn11 does indeed seem to be a valid target, and phenanthroline is probably inhibiting it. But that’s not the best place to be, and a better chemical probe would remove a lot of uncertainty. And I think referring to such compounds as “XXX inhibitors”, as if there were any selectivity involved, can be troublesome for future readers who don’t bother to dig into the literature. It would, for example, be quite interesting to repeat the work in this paper using the compound described in this recent paper, whose authors were searching for a better Rpn11 inhibitor (something better, more believable, and more useful, in other words, than phenanthroline). But that has already been done, by the very discoverers of this new probe: here’s a writeup of its effects on myeloma cells. In this work, the compound was tested beforehand against a number of other metalloproteases to try to get a more selective effect, and I have to regard it as a much more convincing demonstration. Choose your probes well.