New Antibody Data Indicate Tough Battle for RNAi Therapeutic PCSK9 Approach

The past week has been a busy one for the hypercholesterolemia field. New clinical data for PCSK9-targeting monoclonal antibodies from Amgen and Regeneron/Sanofi-Aventis were presented at the 2012 American College of Cardiology meeting, while ISIS/Sanofi-Aventis separately presented long-term efficacy, and especially safety data on the ApoB antisense compound mipomersen (aka KYNAMRO) for which marketing applications have now been submitted in Europe and the US. As hypercholesterolemia also represents a significant medical and commercial opportunity for RNAi Therapeutics due to the advances in knocking down genes in the liver, I will briefly summarize the new data and discuss some of the implications for RNAi approaches, including the phase I candidate ALN-PCS02 by Alnylam.

If there ever were doubts as to the commercial attractiveness of PCSK9 as a target for treating high cholesterol, last week dispelled them all. When it seems that all the thought leaders in the field, including famed cardiologist Steven Nissen, otherwise known for his ultra-critical views of certain medicines, and Wall Street (expected annual sales as a class of up to $20B thrown out) hail the data as proof that PCSK9 will be the final nail in the hypercholesterolemia coffin and thus throw their support behind the class, the commercial success of PCSK9-targeting therapeutics seems a foregone conclusion.

Although pretty much every large pharmaceutical company sports a PCSK9-targeting antibody, the excitement this time centered around the phase II data of Regeneron’s/Sanofi-Aventis’ REGN727 and phase I data of Amgen’s AMG145.

The REGN727 phase II 8-12 week multi-dose studies were conducted in close to 300 patients with elevated ‘bad’ LDL-cholesterol (LDLc) on statins. Depending on the amount and schedule of antibody administered, mean LDLc reductions (note: it is not entirely clear to me whether this refers to the LDLc reduction over time, or the peak LDLc reductions) of 40-73% were achieved. Interestingly, one press report said that ‘727 suffered from relatively short-lived activity such that subcutaneous administrations every 2 weeks would be required.

The phase I studies with ‘727, just published in NEJM (Stein et al.) indicated that the apparent rebound effect after 2 weeks may be due to the concomitant use of statins and that without statins, 50-55% persistent LDLc reductions can be achieved with close to 4-week dosing intervals. This is relevant for example for the statin-intolerant population. With statins, that type of persistent knockdown (more on the higher end of that range) would probably require every 2 week dosing.

AMG145 seemed to attract even more excitement than '727. In the 6 to 8-week phase I studies in around 100 patients taking statins, mean LDLc reductions of 63-75% were observed. These numbers indicate more potent, and likely more persistent LDLc lowering compared to REGN727, although without having seen the LDLc response curves over time, it is difficult to conclude that for sure.

Route of administration and dosing frequency are often cited as important competitive criteria for this class of drugs. In the case of ALN-PCS02 which is enabled by Tekmira’s SNALP technology, the current data indicate that a range of 40-55% persistent LDLc reductions with intravenous dosing every 4 weeks are conceivable. This, however, would require further improvements in potency over the 0.25mg/kg dose level which was the highest dose for which data was presented by Alnylam at their early January update. For this indication, it will also be important to wean the SNALP formulation off the transient immune suppression currently used. This would not only address criticisms that such immune suppressions carry risks, but it also would likely get rid of the observed PCSK9 rebound effect that was apparently linked to it and would thus contribute to prolonging the efficacy.

Of course, an important wildcard in the competition between MAbs and RNAi pertains to the safety of each drug candidate, especially when they are used long-term. The monoclonal antibodies seem to perform quite well in this regard in the short-term, albeit multi-dose studies. RNAi, however, has the theoretical advantage in that it does not involve the formation of antibody-PCSK9 complexes which could eventually have an impact on both long-term safety and efficacy/dosing frequency.

In retrospect, RNAi has picked a tough battle here. As a target that acts extracellularly, readily accessible from the blood, it seems an ideal target for monoclonal antibodies and thus falls outside the undruggable target space that currently still drives the interest of large pharmaceutical companies in RNAi Therapeutic development. On the other hand, the PCSK9 field is populated with antibody approaches, and should there be antibody-specific class adverse effects, RNAi could be the last one standing (note: the antisense candidates by ISIS and Santaris have already dropped out of the race, at least for now), something that one of the few Big Pharma/Biotech that has not yet invested in a PCSK9 candidate may value from a strategic point of view alone (RE partnering).

Meanwhile, the mipomersen extension study data presented this week suggest that the high drop-out rates and liver fat accumulations* (both possibly linked to some extent) at modest ~30% LDLc reductions make it a niche drug for the very small indication of homozygous familial hypercholesterolemia in the US and Europe, and possibly severe hypercholesterolemia in Europe. Although ISIS continues to claim that the liver fat accumulations normalize after reaching maximum median increases of +13% over baseline at week 52 (note: according to the mipo-related Visser et al. 2009 study, 5.6% absolute intrahepatic triglyceride contents are considered the upper-limit of normal), the absolute values would call for caution, especially when there are outliers in whom almost half the liver consists of fat. In the related conference call, ISIS further claimed that the apparent normalizations were not the result of the patients that discontinued mipomersen (possibly due to high liver fat contents). However, since ISIS has a history of erring on the side of optimism when it comes to mipomersen safety and tolerability, I’d like to see the liver fat-related sub-group analyses that take into account the discontinuations, dose reductions and interruptions.

The liver fat data may also mean that ApoB is unlikely to be a suitable stand-alone target for an RNAi Therapeutic, especially if even more pronounced ApoB knockdowns can be achieved than with mipomersen. Consequently, in order to exploit the therapeutic utility of ApoB as the critical protein of atherogenic lipoproteins, an ApoB-targeting RNAi Therapeutic should involve at least one other target gene which will also reduce liver fat content. It remains to be seen whether ApoC-III can be that target.

* Note: Despite the caution expressed in this article, I should add that there is general controversy about the relevance of simply elevated liver fat (NAFLD) in the absence of inflammation.