Five months ago we introduced PERL101, one of our lead compounds for Niemann-Pick Type C (NPC) disease, in Part One. Here in Part Two we pick up the thread. On this journey we’ve progressed from nematodes to patient cells to mice with an unoptimized primary screening hit, bypassing hit-to-lead and shortening lead optimization. As we begin breeding up NPC1 knockout mice for PERL101 efficacy studies, let’s take a few moments to consider how we got here.
It all started with the generation of a NPC nematode model using CRISPR in the end of 2014, followed by validation of a NPC nematode screen in early 2015. Our CRISPR-generated nematode model has a null mutation in ncr-1, the nematode ortholog of the human NPC1 gene; and it has two developmental phenotypes caused by the absence of cholesterol. These phenotypes were previously described in null mutants from a chemical mutagenesis screen over 15 years ago: delayed onset to adulthood and smaller brood sizes.
The worms on the left are ncr-1 mutants grown in the presence of cholesterol. The worms on the right are ncr-1 mutants grown in the absence of cholesterol. Basically, lots of sick juveniles. That is to say NPC is a childhood developmental disorder in nematodes, too.
In the first five months of 2015, we were able to reverse nematode NPC disease with small molecules in a 50,000-compound high-throughput screening campaign. By August 2015 we had 100 fully validated hits. At least a dozen of these 100 hits are also active in NPC patient cells. We named the first compound in this series PERL101.
PERL101 nearly doubles the total well area occupied by worms. The pixel increase is attributable to thicker individuals, i.e., larvae maturing to egg-bearing adults. In more quantitative followup egg-counting experiments using individual worms on plates, we found that PERL101 nearly doubles brood size relative to controls, either with or without cholesterol. Because PERL101 rescues in the absence of NPC1 protein function, we refer to it as a pharmacological bypass suppressor.
Next we ran PERL101 through the first gauntlet of in vitro drug metabolism assays and in vivo pharmacokinetics/PK experiments. Despite the fact that nematodes are invertebrates, it quickly became apparent that PERL101 has very favorable PK properties and 100% oral bioavailability in mice. Were these just lucky rolls of the drug discovery dice?
I argued in Part One that compounds with favorable PK and oral bioavailability are selected for in organism-based screens — even in species that don’t appear sufficiently complex — but not in human-cell-based screens, and forget about target-based screens. How many other pharmaceutical properties besides PK and oral absorption can be selected for in a nematode?
A critical attribute for any new first-in-class NPC drug candidate is blood-brain barrier penetrability because neurodegeneration plays a large role in NPC disease. Cyclodextrin, which is currently in a repurposing clinical trial for NPC, does not cross the BBB and is therefore administered by lumbar puncture (aka spinal tap). After a single oral dose at either 20 mpk or 80 mpk, PERL101 levels were measured in the plasma and brain of mice at several time points up to six hours post-administration. Not only does PERL101 permeate the brain, it reaches 1:1 tissue concentrations relative to plasma.
These in vivo data are corroborated by in vitro experiments. A transport and efflux assay using MDCK-MDR1 cells revealed that PERL101 is neither a substrate nor an inhibitor of P-glycoprotein. We also knew from plasma protein binding experiments that PERL101 is only 75% plasma bound, resulting in enough free PERL101 to distribute throughout the mouse. Nematodes obviously don’t have livers, kidneys, stomachs or brains. PERL101 behaves pharmacokinetically as though they did.
But the biggest preclinical concern of all is…toxicity.
Toxicity is the slayer of many an otherwise promising lead compound. For us to show that PERL101 extends the lifespan of NPC1 knockout mice, animals must be dosed daily for weeks and potentially months. After an exploratory 21-day tolerability study in September 2015, we knew that wildtype Balb/c mice could tolerate up to 200 mpk p.o. q.d. for a week before they lost too much body weight and reached humane endpoint (> 20% body weight loss).
In October/November 2015 we began a 90-day maximum tolerated dose (MTD) study in wildtype mice with three dose groups: 120 mpk, 80 mpk and 60 mpk. Here’s a summary plot of the 90-day MTD study, which includes an age-matched vehicle control group.
80 mpk proved to be the maximum tolerated dose, because the 120 mpk dose group was euthanized after 35 days. However, 60 mpk does not appear to be a no-effect dose. PERL101 affects body weight over time with a peculiar dose response. Whereas the 60 mpk dose group displayed decelerated weight gain relative to control mice, the 80 mpk dose group displayed accelerated weight gain relative to control mice. When looking at body weight change at the level of individual mice (orange are PERL101-treated, gray are controls), there are two PERL101-treated outlier mice but the trend toward increased body weight approaches significance, especially in the second month of the study.
So PERL101 is well-tolerated in mice after 3 months of continuous high dosing. Like good PK and oral bioavailability, is tolerability another essential pharmaceutical property that is predictably selected for in a simple animal once the extent of evolutionary conservation hiding in plain view is fully appreciated?
Although the lifespan of a nematode is only 2-3 weeks, the way our drug treatment assays are designed the worms spend a large portion of their life continuously exposed to PERL101, which they imbibe along with their food. By using a whole animal screen with a positive growth readout and by dosing orally over time, we created conditions that selected for stable, tolerable and target-engaging hits that are can be validated in mice without delay or optimization.