Skip to main content

"Me Too" Drugs

Proteasome Inhibitors, Refined

The proteasome is quite the structure. It is the shredding unit of the cell, where no-longer-needed proteins go to be ripped down to their components for recycling, and it’s become a more and more important part of drug discovery over the years. For one thing, all this fashionable targeted protein degradation work is about sending designated proteins off for destruction at will and before their time. But even before that, the opposite process – proteasome inhibition – had made an impact.

That’s because there are cancer cells that (because of their higher metabolism, greater production of various proteins, etc.) are more sensitive to problems with proteasomal function, and will die more quickly when the waste-disposal service is interrupted. The first compound on the market to do this was Velcade (bortezomib), which was approved in 2003 for multiple myeloma. It came out of a small company called Myogenics, then was developed further at Millennium, and was also notable for having a boronic acid head group. Those had been known as enzyme inhibitors for a long time, but were thought by many to be unfit to get all the way through clinical trials (after a number of failed development projects during the 1980s and onward).

Another inhibitor, Kyprolis (carfilzomib) was approved in 2012, and that one has an epoxide as the warhead, which is also a bit unusual (it had its origins in a natural product). In 2015 came Ninlaro (ixazomib), another boronic acid. Still in clinical trials are the epoxide oprozomib, the boronic acid delanzomib, and the unusual beta-lactone marizomib (also known as the marine natural product Salinosporamide A). That last one is notable for having gone from “here’s a new natural product” to “here’s an IND application” in what may be record time. Now, this list might be seen as an egregious example of “me-too” drugs gone crazy. How, you wonder, did the world end up with so many proteasome inhibitors, especially since several of them look so similar to others in the same class?

Because they’re still different. The proteasome is a rather complex structure, and its core is made up of a number of protein subunits that form a broad cylinder. The alpha-subunits are at both ends of this tube of destruction, interacting with separate cap protein assemblies that regulate what gets in, and the beta-subunits contain the proteolytic enzymes that form the cylinder walls (in two layers, yet). That graphic at right is from the abstract of this recent paper, from St. Gallen and Leiden, which is a head-to-head comparison of the different compounds in their activity against these subunits and what the means for multiple myeloma therapy. As you can see, the two epoxide compounds (especially at low dose), shut down only the beta-5 subunit. The three boronic acids, though, are active against both the beta-5 and beta-1 proteins. The epoxide carfilzomib, at higher doses, starts inhibiting the beta-2 subunit as well as the beta-5. And the natural product marizomib hits all three of the proteolytic subunits simultaneously. As you’d imagine, the stress on the cells increases as you shut down more of these, but (as the paper rightly notes) the different functions of these subunits (the three in the inner ring and the three in the outer, not shown in the schematic) are not well understood at all.

It had already been known that beta-5 inhibition by itself was the weakest mode of action, and sufficient to take out only the most sensitive cells. The beta-5/beta-1 and beta-5/beta-2 combinations were known to be more effective, but this paper has evidence from cell studies that the latter is a better bet – in fact, that the amount of beta-2 inhibition added to the beta-5 activity is the single biggest factor in differentiating the drugs. What’s more, allosteric effects (what happens to the inhibition profile of the beta-2 subunit after the beta-5 is already inhibited) should make it a better idea to use two separate specific inhibitor compounds rather than one bispecific one, since you could titrate in the beta-2 inhibition much better that way.

This means that higher doses of carfilzomib are perhaps the best mode of treatment among the existing compounds, but the complication is the possibility of cardiac toxicity when you do that, since you’re hitting both subunits at the same ratio. The heart seems to have fewer active proteasomes in its cells than almost any other tissue, which gives you less of a margin for error in shutting them down. I’m unaware of a compound that has only beta-2 activity, but if that’s even possible, it could well be a valuable drug candidate in combination with one of the specific beta-5 inhibitors. Update: see the comments! Meanwhile, marizomib only inhibits beta-2 in a fairly narrow concentration range compared to its other activities, and if this paper is right, that means that it could have an uncertain future in its clinical results. . .

13 comments on “Proteasome Inhibitors, Refined”

  1. Larry Dick says:

    Great post! Bortezomib was discovered at the biotech named “MyoGenics” (not Myogenetics). MyoGenics changed its name to ProScript around about 1995. The capital letters in the middle of those names was the fashion of the times.

    1. Derek Lowe says:

      Just fixed it – thanks!

  2. cynical1 says:

    You wrote “Meanwhile, marizomib only inhibits beta-2 in a fairly narrow concentration range compared to its other activities…….” I am not sure what you mean by that because you also wrote “And the natural product marizomib hits all three of the proteolytic subunits simultaneously”.

    The beta subunits in the proteasome are proteases. My experience with (competitive) protease inhibitors is that there is not a “range of concentrations” that inhibit the enzyme. The concentration above the IC90 has pretty much shut down the enzyme. Does this compound stop inhibiting the enzyme at higher concentrations or is it just much less potent against the beta-2 subunit? Or does it lose activity against beta-2 once the beta-5 subunit is fully inhibited to which you also elude?

    I am missing something.

  3. matt says:

    You write that proteasome inhibitors were thought to be useful against cancer because cancer cells for various reasons were more sensitive to proteasome disruption. You write that beta-5 inhibition was thought to only take out the most sensitive cells. Doesn’t that make them ideal, then? Or does it turn out to need to take out a lot more than the most sensitive cells to really hit cancer, and at what point is it really killing all cells and the cancer cell PK (either from higher vascularization or other means) is just causing it to kill cancer cells slightly sooner or more frequently?

    1. Derek Lowe says:

      A better way of phrasing it might have been “take out the more sensitive cancer cells”, I think. It would seem that only hitting them through that mechanism will allow many to survive. . .

  4. Izidor says:

    Hi all,
    Very recently, beta2 and beta2i (i stands for immunoproteasome, an isoform of proteasome) selective inhibitors were in fact published:


  5. Gerjan says:


    Actually selective inhibitors for each subunit of the immuno and constitutive proteasome have been discovered, see

    1. MC says:

      And have a look at what Kezar are doing

  6. NJBiologist says:

    Am I reading that right–the boronic acids all have the same subunit preference, the sole beta-lactam has its own subunit preference, but the epoxides have at least two preference patterns?

    Are the non-epoxide parts of those molecules doing some targeting?

    1. Commentor0451 says:

      The peptide side chain substituents are doing most of the targeting, that is what gives the different proteasome subunits unique substrate preferences in vivo in the first place, after all.

      Also this conversation is massively missing the immunoproteasome subunits Beta1i, Beta2i, and Beta5i, which are expressed by multiple myeloma cells and other immune cells.

      Selective Beta5 or selective Beta5i inhibitors are relatively non-cytotoxic, but ixazomib, bortezomib, and carfilzomib all inhibit both Beta5 and Beta5i at relatively similar concentrations which makes them quite cytotoxic. See F Parlati et al. 2009 in Blood for example.

      If the P1 side chain (closest to the warhead) is leucine, you tend to favor Beta5 and Beta1i. Phenylalanine tends to bias towards Beta5i. Canonically Beta2 and Beta2i prefer Arginine (hence the Trypsin-like name for their substrate preference) and Beta1 prefers Aspartic acid (hence Caspase-like).

      1. Commentor0451 says:

        I should mention the paper itself does mention this stuff, not to imply it is missing info on the immunoproteasome subunits or the lack of toxicity of beta5-only targeting.

      2. Derek Lowe says:

        That’s quite true – I left that part of the story out, although the paper does go into it.

  7. bacillus says:

    Regardless of their mode of action, they add to the arsenal of drugs effective against multiple myeloma so much so, that MM has like AIDS become a chronic disease that can now be treated to such an extent that lifespan of patients has been greatly imporved from a median survival time of 5 years post-diagnosis (MM) to possibly 20+ years.

Comments are closed.