Skip to main content

Cardiovascular Disease

CETP Finally Heads to the Trash Heap?

The tale of CETP (cholesteryl ester transfer protein) as a drug target has been long, and convoluted, and expensive, and horrendously disappointing. Pfizer failed, Roche failed, several other companies like BMS who didn’t even get that far failed, and last fall Lilly’s entry failed, too (that link has links to several past blog posts here on the subject – I’ve been writing on the subject for over ten years now).

More details have now emerged about that most recent exercise in futility – and “clinical futility” was exactly what killed the compound off. And it’s quite interesting, in a disturbing way: Lilly’s evacetrapib did just you would want a CETP inhibitor, or any lipid-targeting cardiovascular drug, to do. It raised HDL by 130% compared to placebo, and lowered LDL by 35%. But the cardiovascular outcomes (MI, angina, stroke, and so on) were absolutely identical between the treatment group and the placebo group. At one point, if you’d asked cardiologists to predict the effects of a compound that affected cholesterol levels in this way, you’d have gotten some pretty enthusiastic guesses. But not now.

In the end, cholesterol, lipoproteins, and all the rest of the things you can get from blood work are merely surrogates for what really matters, which is how many people end up in the hospital and how many die. There’s a pretty good correlation between those numbers and low total cholesterol, and it’s widely believed that there’s a correlation with low LDL as well. And there’s evidence that high HDL is protective, too, thus all the talk of “good cholesterol” and “bad cholesterol”. But if you can raise HDL and lower LDL and still have no effect at all, then just how good are these markers?

There are several possibilities: perhaps the markers are OK, but CETP inhibitors do something else that cancels out the beneficial effects that they’re signaling. That was the thought with Pfizer’s torcetrapib, where the treatment group actually had higher mortality than the control group. Effects on blood pressure were seen, for example, and in this Lilly trial there was a small (but statistically discernable) rise in blood pressure as well. But it doesn’t seem to be enough to completely cancel out what the beneficial effects should be of more than doubling HDL, that’s for sure. It might also be that raising HDL and lowering LDL through this specific mechanism somehow doesn’t count, that it comes in the wrong place in the grand lipid-handling scheme to exert a beneficial effect.

Another possibility is that the markers are OK, as far as they go, but we’re not looking closely enough. HDL and LDL can be subdivided into further lipoprotein fractions, and maybe at that level of detail we can see the good good cholesterol as opposed the ordinary good cholesterol, etc. To the best of my knowledge, though, people have been digging around in the Pfizer and Roche data to see if they can make something of this idea, and I haven’t seen anything dramatic. But human lipid handling is extremely complex, and you can plausibly appeal to that at any point when things go wrong.

But that brings up a third possibility, that lipidology is complex enough to have fooled us completely, and that HDL and LDL are not, in fact, good markers at all. There’s a lot of room to argue against that – statins, for example, really do seem to have beneficial real-world outcomes over multiyear dosing. The counterargument is that not all of that is due to lowering LDL per se, that there are other effects (both on- and off-target) that we’re not taking account of. On the HDL side, you have human genetic mutant populations who have abnormally high levels, and do indeed seem to have protection from cardiovascular disease. These two lipoproteins have emerged as markers, then, for some pretty good reasons – but the CETP story is clear evidence that we’re missing something, or several things.

Several large companies have poured hundreds of millions of dollars apiece down the drain in finding this out for us. And Merck – last anyone heard – is supposedly pressing on. But man, you’d have to think that no one over there has a good feeling about this, with the whole CETP landscape lit up by piles of burning cash and striped with the shadows thrown by mounds of clinical debris. It’s a wasteland out there, badlands that have made every single company that’s ever entered the territory regret their actions deeply. And that’s where you’ll find Merck. For how much longer, one wonders?

Related Content

41 comments on “CETP Finally Heads to the Trash Heap?”

  1. Curious Wavefunction says:

    Another nail in the coffin of reductionism.

    1. Peter S. Shenkin says:

      You seem to have left out the obligatory smiley face. “A theory should be as simple as possible, but not simpler.” Reductionism still works, but you have to know what the pieces are and how they work. All this proves is that we don’t. (And I think you know this, but for the rare gentle reader who might not….)

  2. Carl 'SAI' Mitchell says:

    The other side of the third possibility is that LDL and HDL levels are indicators of some underlying process, and not causative at all. So CETP inhibitors might just be changing the indicator while doing nothing to the actual cause of the heart disease, while statins are doing something to the underlying cause, which in turn changes cholesterol levels. IE the old “correlation is not causation” coming to bite us yet again.

    1. Bagnar says:

      I’m not an expert in this specific area, but would it be also possible that there are two or more different conjugated process ?
      Like to open some safes, you need to turn two keys at the same time, to be efficient, a possible treatment should activate two target at the same time.

  3. watcher says:

    Two other possibilities independent of the markers themselves:
    1) LDL/HDL levels and ratios are important if intrinsically generated (eg biologically), not so much if externally modified (eg with drugs)
    2) the favorable / unfavorable underlying cause of LDL/HDL ratios and/or levels is genetic, eg is something that has an influence in potential outcome for ones entire life, not just for a period of time of a few months or years on a drug that manipulates the levels.
    Personally, I believe that “assigned” markers for risk may not be a causative, and the effects of statins is from a combination of activities including general anti inflammatory. And as mush as we would like to think we know, we know much less than we do about these matters.

  4. Bob Hanky says:

    These results do not bode well for Repatha and Praluent…

    1. Mike says:

      How does the clinical failure of CETP inhibitors affect the value of the PCSK9 inhibitors evolocumab (Repatha) and alirocumab (Praluent)? These two drugs work through a different mechanism of action, have been shown to be clinically effective, and are FDA-approved.

      1. Bob Hanky says:

        Both mechanisms assume that certain types of cholesterol are bad or good, both mechanisms assume statins improve cardio outcomes via cholesterol.

        1. milkshake says:

          statin preventive action looks increasingly more like off-target effect on inflammation, somewhere upstream of a signal that triggers formation of fragile plaques (through foam cell formation from the invading macrophages. )

  5. Mark Thorson says:

    The consensus seems to be that the initial event in cardiovascular disease is endothelial dysfunction. Cholesterol-rich plaque formation is way downstream of that, and it may be a defensive reaction to injury caused by earlier stages of the primary disease process. Endothelial dysfunction is also strongly implicated in type 2 diabetes, and a growing body of evidence suggests its involved with Alzheimer’s disease. Pharmaceutical approaches to endothelial dysfunction have great potential and we already have a few, but we need more and better ones. None of the drugs which affect endothelial function were developed for that purpose.

    1. Anon2 says:

      Ivor Cummins/Fat Emperor has been harping on this for years now. Easily accessible/palatable information for a lay audience.

    2. John Smith says:

      Watch out! You “almost” eluded to the real cause there, but you better not piss off the pill pushers. Yes, UN-DIAGNOSED DIABETES is the REAL driver here folks.

      The whole concept of trying to tweak/change one’s HDL via a pill is virtually the same as moving the hands on an analog clock in an attempt at time travel. Yes, the clock will show a different time, but temporally you haven’t really gone anywhere. HDL & LDL are way too down stream of the real drivers, so they can only be used as latent indicators. So trying to artificially change them will never fix the underlying cause (treat the cause, not the symptoms).

      As Anon2 stated, go look at the links he sited, or specifically watch this video of Ivor Cummins trying to explain it so that even an egg-head “should” be able to understand it here:

      BTW, Ivor has approached this whole CAD & CVD problem from an engineering perspective (i.e. he’s not just some uneducated idiot spouting the latest health trend) which tends to be very pragmatic and driven by testable results and research.

  6. Mol Biologist says:

    Thank you Derek again, for bringing my favorite topic.
    “Did You Lose the Keys Here?” “No, But the Light Is Much Better Here”.
    Again, it is hard to admit that limited resource is necessary but not sufficient part of a discovery. Why all great chemists did synthesize new compounds and never ask highly related biological question. Did Akira Endo discovered main function of ML-236B or the substance was doing something else in fungus? Conjugated process is an interlinkage of response to environmental situation. Cholesterol-rich plaque formation is inefficiency one system vs other or unbalanced variation.

    Will Repatha and Praluent be good drugs or not is the next level question. For my personal view to know the answer you have to go back to the roots.

    1. Diver dude says:

      No, the only way to find out if these drugs work is to try them in the clinic. Everything else is just guessing. Highly paid and extremely expensive guessing, but guessing nonetheless.

      “In very large numbers of homo, veritas”

      1. Mol Biologist says:

        Nope, you need to be hunger in paradise. Same things work for sport, is not it.
        The Gold Mine Effect: Crack the Secrets of High Performance. Read it.

        Same Rules for science…..discovery.

  7. Anon says:

    There’s nothing wrong with hubris – as long as you avoid it like the plague.

  8. Peter Ellis says:

    Option (4): LDL and HDL levels are perfectly good indicators of cardiac risk, but they’re a downstream process. Growing a beard is a pretty good (not infallible) indicator of maleness. Shaving it off won’t change your sex.

    1. Mol Biologist says:

      Option (5). You right it is sex effect. This study suggests that increased risk of CAD in FH is not solely due to elevated LDL cholesterol levels and demonstrates a sex-specific lipoprotein influence on CAD in a large sample of FH patients carrying the same LDL receptor gene defect.

    2. Nick K says:

      Excellent analogy! I’m stealing it!

    3. Paramus says:

      Lilly has invested millions in shaving beards! The same analogy applies to beta-amyloid and Alzheimer’s

  9. Barry says:

    in 2003, a very small clinical trial showed that ApoA-1 Milano–reverses plaque formation in humans, without raising the HDL. I see that–after Torcetrapib failed–Pfizer has bought the project(now called “ETC216”, and Esperion) but then spun it out again in 2008. Perhaps they saw no market for a chronic CV therapy without oral bioavailability, and didn’t want to develop it for an acute indication? Now renamed MDCO-216, it was being pursued by The Medicines Company. But if they’ve gotten it into the clinic, it hasn’t been reported.

  10. Barry says:

    ore 2. Epidemiology and Prevention of CV Disease: Physiology, Pharmacology and Lifestyle
    Session Title: Lipid-Lowering Trials
    Abstract 9907: A Single Infusion of MDCO-216 (ApoA-1 Milano/POPC) Induces Marked Changes on the Lipid Profile
    David G Kallend1; Alain Bobillier1; S. Eralp Belibas2; Herman Kempen1; Jakobus Burggraaf3; Joannes Reijers3; Matthijs Moerland3; Peter L Wijngaard1
    + Author Affiliations

    1Rsch & Development, The Medicines Company, Zurich, Switzerland
    2Rsch & Development, The Medicines Company, Parsippany, NJ
    3Rsch, Cntr for Human Drug Rsch, Leiden, Netherlands

    Introduction: Human carriers of the apolipoprotein A-1 Milano variant have a reduced incidence of cardiovascular disease. Regression of atherosclerosis was observed in a prior Phase II intra-vascular ultrasound study. MDCO-216 (ApoA-1 Milano/POPC) manufactured by a more efficient process is being developed to reduce cardiovascular events in acute coronary syndrome patients. We report the results of a Phase I single ascending dose study.

    Methods: After signing informed consent and meeting all eligibility criteria, 24 healthy volunteers and 24 patients with documented CAD received a 2-hour infusion of MDCO-216 in a randomised, placebo controlled, single ascending dose study. 5 cohorts of healthy volunteers and 4 cohorts of CAD patients received doses ranging from 5 – 40 mg/kg. Subjects were followed for 30 days and returned throughout the study for safety assessments.

    Results: In both healthy volunteers and stable CAD patients, dose dependent increases in ApoA-1, phospholipid and prebeta-1 HDL and decreases in Apo E were observed. Prominent sustained increases in triglyceride and decreases in HDL-C occurred at doses above 20 mg/kg in healthy volunteers and patients with CAD. In both subject populations, profound increases in ABCA1 mediated cholesterol efflux were observed. Other lipid and lipoprotein parameters were generally unchanged with increases in HDL particle size and a shift to larger HDL particles. MDCO-216 was well tolerated with no serious adverse events and no other significant adverse safety findings including laboratory parameters, vital signs and ECGs. The most common adverse events were headache and fatigue. No dose dependent effect on any safety parameter was noted and there was no difference in safety parameters between healthy volunteers and patients with CAD.

    Conclusions: A single infusion of MDCO-216 modulated lipid parameters, profoundly increased ABCA1 mediated cholesterol efflux and was well tolerated. These single dose data are consistent with the lipid profile of carriers of the Apo A-1 Milano variant and support further development of this agent as a disease-modifying treatment for reducing atherosclerotic disease and subsequent cardiovascular events.

  11. Natural says:

    “Exercise increases HDL and lowers LDL, so if we could only do those things with a pill we’d be set!”

    Ignoring the millions of other changes to the body caused by exercise.

    1. Mark Thorson says:

      Exercise is the most reliable thing you can do to improve your endothelial function, so it’s upstream of a whole cascade of adverse consequences. Now, if you could develop a pill that can do the same thing for endothelial function that exercise does, that could be a blockbuster like nothing we’ve ever seen before. The indications for use are any one of fat, smoker, or over 40.

      1. Mol Biologist says:

        The Light Is Much Better Here. However, I will vote for unexpected low profile drug candidate vs miracle of blockbuster.
        Because exploring same targets again and again will not change the landscape lit up by piles of burning cash and striped with the shadows thrown by mounds of clinical debris.

      2. KRL says:

        What drugs have been developed to improve endothelial function?

  12. Rock says:

    ” It might also be that raising HDL and lowering LDL through this specific mechanism somehow doesn’t count, that it comes in the wrong place in the grand lipid-handling scheme to exert a beneficial effect.”

    I will repeat the same comments I have made to virtually every one of your 10 previous posts on CETP. In my view, the quote above is the most likely explanation. HDL, VLDL and LDL work in concert through CETP to shuttle cholesterol to the liver. By blocking CETP, you back up the process, stuffing more cholesterol into HDL particles so there is less in LDL. Why would anyone think that should work? Perhaps the field has advanced further since I was following it, but at the time, the number of HDL particles did not increase with CETP inhibition, unlike what happens with exercise. It is like bringing a larger dump truck to an excavation site but the dumping mechanism doesn’t work. You can load it up with as much dirt as you want, but it is just going to sit there.

    1. Mol Biologist says:

      Agreed. Simple and Brilliant explanation. We are more like a fungus. The demand must be resolved at key point not at the last stop.

    2. David Borhani says:

      I agree (having worked on HDL some years ago). It’s the flux from periphery to the liver that’s important, and it’s no great surprise that backing up the works isn’t helpful.

    3. tangent says:

      What are the likely reasons for the LDL/HDL variation in the population that causes the misleading correlation?

    4. Kent Kemmish says:

      Rock can you opine on the Milano approach?

  13. Piero says:

    LDL and HDL could be just that, markers, effect and not the cause.
    If you have a pressure meter on a reactor signaling dangerous pressure, you can take a screwdriver and reset the meter to safe values, this wouldn’t relief pressure in the reactor 😉

  14. Andre says:

    I seem to be missing something here. Rare autosomal dominant mutations in the human CETP gene ( cause reduced function of CETP and have been linked to accelerated atherosclerosis (Zhong et al. 1996 JCI 97:2917-2923). So why would you want to develop an pharmacological inhibitor of CETP? Wouldn’t you just accelerate atherosclerosis in patients taking a CETP inhibitor? If anything you may want to have a drug that increases CETP levels as this may be beneficial in preventing atherosclerosis.

    1. tangent says:

      Yeah, aren’t there more papers at this point saying loss-of-CEPT mutations are bad for CV risk than saying it’s good? And that various other causes for increased HDL are bad? All of these drugs seem like very expensive zombie-dinosaurs from an era when people thought they would work.

      Or if I grasped the whole literature would the weight of evidence still be in the other direction?

  15. Andre says:

    I just noticed that the genomes of mice and rats lack Cetp genes. Seems that life for rodents is just fine without Cetp. This also raises the question of the validity of animal models used for preclinical development of CETP inhibitors….

  16. Among kitchen gadget stores, this site offers many Norpro products which might be not only useful but attractive.
    The biggest glitch is pressing an important when dealing using a computer telephone call.
    Milwaukee battery warranty It’s frustrating when I should make an appointment and have to
    constantly reset the telephone.

    If you make use of a cordless expandable phone system with Vo – IP, but it completely eliminates the
    need for a fool, but the wires. They may file a fraudulent tax return using your details.

  17. Kent Kemmish says:

    What happened with the Milano allele work? I haven’t follow the atheroma reduction part of SENS since 2008. It seemed smart and promising.

Comments are closed.