RG-101 for HCV: Give Me 8mg/kg

Last October, Regulus Therapeutics announced stunning data from an ongoing trial of anti-miR122 RG-101 for the treatment of HCV infection.  Accordingly, 6 out of 14 patients administered a single injection of 2mg/kg of the GalNAc-conjugated antisense oligonucleotide dropped to what the company referred to as ‘below the level of quantitation’ (BLOQ) by day 30 and remained there as of the update (day 57 for those treated first in the study).

*BLOQ is not the same as 'undetectable'.  It is 'undetectable' for x number of weeks that technically defines a cure.

2mg/kg data looking ever better

The more I look at the data (here for the corresponding dataset), the more intrigued I get.  For starters, the 2mg/kg was only the starting dose in the patient cohort.  Moreover, a patient that had been more recently treated and had reached only 29 day in late October seemed to be on his way to BLOQ.  In fact, after his titers had dropped to below ~1000 IU/ml, viral decline seemed to accelerate as if a threshold had been reached. 

Overall, it seemed that as long as you got below ~1000 IU/ml, the virus would not rebound (at least until day 57).  Accordingly, the 5 patients that rebounded got close (1000-10000 IU/ml), but never below the apparent threshold.  The lack of overlap between the individual knockdown curves also suggest that viral rebound is not the result of stochastic escape mutations, as is often the case in HCV, but simply due to diminishing drug concentrations.

This leaves two patients which as of the update were still slowly declining as if the threshold rule was counterbalanced by declining oligonucleotide concentrations in the liver after day 30.

In summary:

Everybody responded irrespective of genotype and starting viral titers, of these…

…6/14 BLOQ;

…3/14 undecided;

…5/14 nadir of 1000-10000 IU/ml, then rebound.

A one-shot cure?

In the current marketplace, comprising of ~12M untreated HCV patients in the top 7 pharmaceutical markets plus hundreds millions more in the rest of the world, HCV treatment success has less become a function of viral cure rates achieved in clinical trials, but more one of compliance with the pill regimens.  Therefore, a therapy which would shorten current treatment regimens (say to 4 instead of 8-12 weeks), or even involve just one or two subcutaneous administrations in the physician’s office as part of routine checkup visits, would have great pharmacoeconomic impact.

The company is currently crunching the numbers for the next higher, 4mg/kg cohort.  In general, with RNAi and antisense therapeutics, what you see with increased doses is a slight acceleration of gene knockdown (thereby shifting the knockdown curves down), and, likely of more importance in this case, extended duration of drug activity.

The latter is supported by animal, including non-human primate data that showed that liver concentrations of RG-101 increase linearly with dose from 1-10mg/kg.  This means that for the 3 undecided patients in the 2mg/kg cohort, a doubling of the drug concentration could have easily pushed them BLOQ.  

Alternatively, a second dose of 2mg/kg on day 29 would have achieved that as also supported by animal data showing that viral nadir drops further when a second dose is given (on the left is an example from the challenging chimeric mouse model).

I am less sure whether the other 5 patients would have gotten there with 4mg/kg, but would be much more sanguine about it with a dose of 8mg/kg.

A mistake not to escalate to 8mg/kg

Unfortunately, Regulus Therapeutics stated last year that based on biomarker data which measures the activity of human genes naturally regulated by miR-122, 4mg/kg would be the highest dose to be tested in HCV.  This is because the biomarker response seemed to have saturated already by 2-4mg/kg.  This despite of the fact that they had established in the same phase I study that 8mg/kg is also safe and well tolerated in non-infected healthy volunteers.

I clearly think that stopping at 4mg/kg is a mistake.  First, I highly doubt that at least in terms of duration of biomarker activity, they did not see improvements with higher doses (show me the data).  Second, it is wrong in my opinion to assume that drug saturation with regard to microRNA activity against endogenously expressed genes is a good reflection of microRNA activity in the context of viral replication.  In this case, the rationale is further in question because of the special way miR-122 acts in HCV replication (increases, not decreases activity).

Tempering expectations

As we await the 4mg/kg data and an update on the 2mg/kg cohort in ‘early February’ according to recent company guidance, I should caution that the above scenario suggesting a one-shot monotherapy cure is within reach is obviously an unbelievably exciting best-case scenario.  There is a lot which could make the market react adversely to the upcoming data.  This includes emerging side effects, and responses that could even be worse than the 2mg/kg data, e.g. if some patients do not respond at all or if the company is right in their biomarker assumptions which due to chance could make the 4mg/kg data look worse numerically. 

Therefore, my base case for market-neutral data is that all patients will have to respond, and that the extent of the response will be similar to that seen with the 2mg/kg dose.  But then again, I’m looking at the data and keep thinking to myself…’8mg/kg might be the one-shot cure that could transform HCV treatment’.  In any case, the company has little to lose from dose escalating to 8mg/kg.