Dicerna Exploits HBV Biology For Less Frequent RNAi Dosing

Despite recent findings by Arrowhead Pharmaceuticals suggesting that finite treatment with an RNAi medicine might be possible in the search for a (functional) cure of HBV, less frequent dosing is always a plus.  It now seems that fresh HBV entrant Dicerna may have stumbled across a way to reduce dosing frequency and at the same time uncovered an intriguing piece of HBV biology involving the mysterious HBV X protein (HBx).

Search for perfect RNAi target site uncovers important HBV biology

Gene knockdown trigger selection has become critical in a highly competitive field featuring not only Arrowhead and Dicerna, but also Arbutus, Ionis (along with partner GSK), and the Alnylam-Vir alliance in the Western World alone.  This is because ‘by rotten luck’, Arrowhead found that an RNAi trigger that was designed to hit all the HBV transcript did not take into account the absence of the target site on transcripts that derive from host chromosomally integrated HBV.    

The finding that in many patients, e-antigen negative patients in particular, most viral transcripts derive from integrated HBV, has to be considered the most important new discovery in HBV biology and  disease progression since the finding of NTCP as the viral entry receptor.

This has established the selection of a target site upstream of the integration breakpoint as the new industry standard.  Little did we know that the search of the perfect RNAi target site should yield another fundamental insight into HBV biology that could also be therapeutically valuable and provide for some important competitive differentiation for late entrant Dicerna.     

HBx-sparing RNAi trigger has dramatic impact on HBV core protein localization

When scanning the HBV genome for target sites, Dicerna found that an RNAi trigger that hits all HBV transcripts except for HBx was associated with considerably more sustained gene silencing following a single dose (>>8 weeks in a mouse model where GalNAc-RNAi is generally less long-lived compared to humans) compared to an RNAi trigger hitting all HBV transcripts without exception (~3 weeks until knockdown was considerably lessened). 

Further underlining the longevity of the effect, Arbutus’ LNP-based formulation was only able to produce 1 week of good gene silencing in the same (HDI; slide 6) mouse model.

Interestingly, this effect was correlated with striking differences in the cellular localization of the HBV core protein: when all gene products were equally targeted, core protein was predominantly in the nucleus.  By contrast, core protein was almost exclusively cytoplasmic when HBx was spared.  Since phosphorylation of the C-terminal domain is known to be important for HBx nuclear trafficking, it was speculated that HBx may bind in that region and thus mask the nuclear localization signal.  This would be useful during the late stages of a viral infection when core protein would be needed for pregenomic RNA packaging and ultimately viral release.

The reason why gene silencing would be shorter lived when core was in the nucleus is likely due to its stimulating activity on transcription from the HBV cccDNA which would increase transcript dynamics and thus more quickly dilute out the anti-HBV-loaded RNAi machinery in the cytoplasm.

Indeed, in a non-cccDNA-dependent mouse model (e.g. more reminiscent of e-antigen negative patients), this beneficial effect was not observed.  

In summary, while the mechanistic basis of the differential silencing phenomenon remains to be fully fleshed out by further experimentation, Dicerna has found a way to quite dramatically extend gene silencing and thus decrease drug administration frequency.  This is a useful competitive feature in antivirals in general where drug resistance due to missed doses is always a concern and also simply due to patient convenience.

Disclosure: long DRNA.

Gene Knockdown in Disease Involving Gene Expression Throughout Brain Reported

Today, Ionis issued a press release revelaing that their drug candidate for Huntington’s Disease was able to knock down the huntingtin target gene in a dose-dependent manner.  This is the first clinical demonstration that single-stranded phosphorothioate antisense technology cannot only engage gene targets, as had been shown in the gain-of-function approach for spinal muscular atrophy (SPINRAZA; slide 58), but that it could do so in a sufficiently robust manner so that a knockdown could be measured.

This, of course, has broad implications for the Ionis antisense platform which is similarly being developed for other RNaseH-based knockdown applications ranging from the rare and severe (e.g. ALS, spinal cerebellar ataxias) to the more common neurodegenerative diseases such as Alzheimer’s and Parkinson’s.

The reason why this feat is remarkable and a great de-risking event for the technology is that in some of the indications, the target gene is expressed and needs to be suppressed more or less throughout the brain.  Assuming Ionis didn’t take brain biopsies, but instead looked for protein expression by taking CSF samples which would +/- give you an average of target gene expression in the entire CNS (a safe assumption), the ability to assess a dose-dependent gene knockdown is a testament to the robustness of the gene knockdown.

The actual numbers, however, remain under wraps as Ionis and partner Roche (which has exercised its option to IONS-HTTRx in the wake of the data) plan to present them more formally through a publication and a conference presentation with key thought leaders in the disease present.  My guess is that peak knockdown is in the 50%+ range which Frank Bennett from Ionis has recently referred to be in the desired knockdown range.

Given that dosing in the study only lasted 3 months in this slowly progressive disease, it is unlikely that actual clinical benefits will be reported from this phase I/IIa study.  But given the so called 'huntingtin knockdown holiday phenomenon' and some remarkable comments from investigators in the study and KOLs, one cannot but hope that we’ll be in for a positive surprise.  

New Arrowhead Data Shows RNAi Ratting Out HBV

When Arrowhead suffered its DPC fiasco a year ago, it got a chance to re-prioritize its pipeline.  It therefore caught my attention that despite the significant theory risk that still persisted, it chose HBV as one of its lead indications.

To wit, knocking down HBV genes promised to re-awaken the host immune system thought to be exhausted in chronic HBV patients.  The HBV surface antigen (HBsAg) had been regarded to be the main culprit for this exhaustion, although other gene products are now thought to play a role, too.

Previous update suggested long road ahead

When Arrowhead left off, clinical data had shown robust HBV gene suppression (~1-2log range) when the RNAi trigger target sites were present (in treatment-naïve e-antigen positive patients).  Curiously, in e-antigen negative patients where many HBV transcripts lack the ARC-520 target sites, a modest, protracted anti-HBV response could be observed.

Unfortunately, data presented this spring at EASL suggested that long, if not chronic RNAi dosing might be necessary since viral rebounds were seen in all but one subject once ARC-520 dosing had been stopped after half a year of treatment.

Chronic dosing, if indeed it can be shown to lead to a reduction of cirrhosis, liver cancer, and death, is conceivable with a subcutaneously administered RNAi agent.  The road, however, to proving such value and broad adoption in the clinic would be harder and longer compared to demonstrating host control of the virus after a finite treatment period.

New update shows host gaining upper hand

Based on the latest data presented yesterday at HepDART, including ~8 months of additional follow-up in the same patients, this may not be the case anymore.  This is because in 50% of the subjects (in 2 out of 3 e-positive and 2 out of 5 e-negative) given ARC-520 along with polymerase inhibitor entecavir, there have been sustained anti-viral responses after the initial rebound. 

In all cases, HBsAg continued trending down until the last time-point reported, and in 3 of these 4 subjects HBsAg was approaching the limit of detection.   Intriguingly, this was accompanied by a modest spike in liver enzymes, consistent with the host immune system regaining the ability to spot and remove infected hepatocytes. 

Of note, the HBsAg declines are more robust (5.0, 3.1, 2.0, and 0.6) and faster than would be expected based on entecavir treatment alone (~0.5log per year).

Since the enzyme elevations generally followed on the heels of the viral rebound, it is possible that treatment cessation and getting HBV to come out of hiding is actually beneficial.  Nevertheless, in the one subject (patient 01-7985) where there was no viral rebound, ALT values still rose.  

Experimenting with treatment schedules will therefore be an important component of future HBV trials.

In the meantime, we can look forward to an update on the same subjects after another few months. Could it possibly be that we are witnessing  the first HBsAg clearances and host control of HBV brought about by RNAi?