Pages

Monday, August 26, 2013

Progress in Predicting the Immune Response to an RNAi Therapeutics

When in 2010 Tekmira decided to prematurely terminate its first clinical RNAi Therapeutics candidate, TKM-ApoB, due to unexpected immune stimulation in a phase I, it was disappointing in a number of ways.  Not only did it destroy TKM-ApoB right there, it also cast doubts on how predictive our preclinical immune stimulation tests really were and whether as a result, other candidates may suffer similar fates.  Apparently, using rodents or even non-human primates for that purpose had not been all that useful either. It is also said that the event served as a red flag to some regulators which wanted the technology being developed more slowly.

Tekmira, however, soon claimed to have discovered the reason why their PBMC-based assay did not pick up on the immunostimulatory potential of TKM-ApoB and reported that they had have found a more predictive test tube assay instead.  Apparently, using heparin during the preparation of PBMCs (sample enriched for certain immune cells from blood) was the culprit and a whole blood-based assay (including things like red cells and albumin) would reflect much more reliably the situation in people.

A nice paper by Coch et al. from the Hartmann group in Bonn, Germany, now provides comprehensive insight into possible causes for false positives and false negatives when using in vitro immune stimulation assays.


Role of serum proteins  

It is known that phosphorothioate oligonucleotides, a particularly widely used chemistry in the antisense field, rapidly bind to proteins in serum.  Given its abundance and sticky nature itself, albumin is thought to be an important target of such oligos.  The group thus found that their binding in whole blood sequestered them from the immune cells such that the immune response was significantly lessened.  By contrast, in the PBMC-based assay, these phosphorothioates stimulated a robust immune response.

This insight is obviously particularly important if the interest was in developing oligonucleotides that are immunostimulatory on purpose, especially for cancer and viral applications (false positive in PBMC).  Whether the finding that in whole blood assay phosphorothioate oligos may not exhibit immune stimulation can be used to conclude that they will be safe is another question as the blood-based assays may not be informative on what is going on in the tissues.   


Role of anticoagulants

Closer to the TKM-ApoB story, another important insight was that the nature of the anticoagulant can have a big impact on the outcome of the assay.  When drawing blood, and especially when employing a whole blood assay, it is important to use an anticoagulant as otherwise standing blood would start to clot. 

EDTA and heparin are standard reagents for this purpose.  EDTA has been recognized that it can distort immune stimulation assays and thus is not used.  The surprise (well not that surprising in hindsight) is that heparin, due to its negative charge, could displace oligos such as RNAi triggers from their delivery vehicles, including liposomal nanoparticles (LNPs).  It is often the delivery agents that bring the oligonucleotides to the immune receptors.  Moreover, nanoparticle formulations such as LNPs can facilitate multivalent interactions as they hold together and present a number of oligonucleotides simultaneously.  This in turn can potentiate the immune signaling.

Accordingly, what probably caused the immunostimulation of TKM-ApoB to be missed is that heparin at some point displaced the ApoB siRNA from the liposomes thus destroying the potentiator effect of SNALPs.  Importantly, replacing heparin with hirudin (think leeches) did not suffer from the same limitation.  Therefore, human whole blood assays with hirudin as the anticoagulants appear to be the simplest and most reliable preclinical innate immune stimulation assay for Oligonucleotide Therapeutics development.


Growing the database

Of course, these insights are only a first, albeit critical step.  The next step is to understand what degree of immune stimulation in the test tube relates to a likely adverse event in the clinic.  For this, it is necessary to go back and forth between the test tube and clinical observations and correlate the clinical phenotype (cytokine production, fever/chill symptoms) with the test tube response. Complicating matters, different persons have individual innate immunostimulatory sensitivities, and depending on the state of the immune system (e.g. existing infection) there will also be intra-patient variability.
   

On a more positive note, not all delivery/RNAi trigger formats are equally prone to stimulating the immune system with LNPs probably being one of the more challenged formats with regard to innate immune stimulation.  Even so, the clinical track record with SNALPs after TKM-ApoB (lowered doses of steroids with ALN-TTR02, no steroids apparently used in the TKM-EBOLA trial) suggests that the new assays are having a positive impact already.

5 comments:

  1. How's Arrowhead doing on this front with their ARC520? How would we know?

    ReplyDelete
  2. RNAi triggers attached to DPCs are quite a bit 2'O-methyl/F modified so the inherent immunostimulatory potential should be very low in general. The ARC520 data have borne this out so far.

    ReplyDelete
  3. Thanks, Dirk.

    Will it matter, in your opinion, that unlike the prototypical DPC, the DPC polymer and the siRNA will be separate, or, as mentioned in Arrowhead's presentation:

    "CDM-masked endosomolytic polymer and siRNA are NOT attached and do NOT interact. Targeted independently to the same cell after co-injection"

    ReplyDelete
  4. Yes, it should matter and should further decrease already relatively low chances of immune activation in blood.

    ReplyDelete
  5. Chances of immune activation in blood even "further decreased" by the two not being attached.

    Thanks, Dirk. Much relieved to read that.

    ReplyDelete