PREDICTION OF HUMAN PHARMACOKINETIC PROFILES OF GLYPH PLATFORM PRODRUGS USING ANIMAL AND HUMAN DATA VIA A SEMI-MECHANISTIC PHARMACOKINETIC MODEL

Background

GlyphAllo™ (SPT-300 or Glyph Allopregnanolone), an oral prodrug of allopregnanolone, is designed to overcome bioavailability limitations of allopregnanolone by shifting absorption toward the lymphatic system, avoiding first-pass hepatic metabolism. GlyphAllo uses the Glyph™ platform, a lymphatic-targeting prodrug technology which reversibly conjugates a drug of interest to a dietary fat molecule using proprietary chemistry, enabling higher oral bioavailability of allopregnanolone. A phase 1 clinical trial (NCT05129865) demonstrated that GlyphAllo dosing in healthy volunteers was generally well tolerated and resulted in therapeutically relevant plasma exposures of allopregnanolone. The objective of this analysis was to develop a semi-mechanistic pharmacokinetic (PK) model to predict human exposures for lymphatic-targeting Glyph prodrugs from preclinical data.

Methods

Allopregnanolone PK profiles were obtained after a single oral dose of GlyphAllo in cynomolgus monkeys and a first-in-human phase 1 study (NCT05129865). A semimechanistic PK model was developed with an absorption compartment, five transit compartments to account for the unique absorption kinetics of dietary lipids, and central and peripheral compartments to describe both the monkey and human profiles. Using allometric scaling, clearance (CL), central volume (V1), peripheral volume (V2), and intercompartmental rate transfer (Q), monkey PK parameters were obtained from a published human intravenous (IV) allopregnanolone population PK model1. To capture the unique absorption kinetics of lymphatic delivery, optimal absorption (ka) and transit rate (ktr) PK parameters were fit to the semi-mechanistic PK model using the monkey allopregnanolone concentration-time profiles. These parameters (ka, ktr) derived from monkey were then combined with human published PK parameters1 to model the profiles measured in the phase 1 study. All modeling building and graphical analyses were done using NLMIXR2 (Version 3.0.2), RStudio (Version 2024.12.0), and R (Version 4.4.2).

Results

The semi-mechanistic PK model with first order absorption along with five transit compartments and central and peripheral compartments had a better fit of cynomolgus monkey allopregnanolone PK profiles compared to other PK models with fewer or no transit compartments. No misspecification in the structural model was observed. The results showed that the absorption and transit rate constants from the monkeys translated well to humans and were able to capture the shift in time to maximum concentration (Tmax) and maximum concentration (Cmax) of allopregnanolone following oral dosing of GlyphAllo. The goodness-of-fit plots showed a high degree of alignment between predicted and observed data for both monkey and human.

Conclusion

We present a modeling approach to predict human exposures of lymphatictargeting Glyph prodrugs by first using human PK data of existing drugs and backtransforming them to animal PK data, then subsequently fitting optimal absorption and transit compartment parameters in cynomolgus monkey to capture complex lymphatic absorption kinetics. These absorption parameters from cynomolgus monkeys, along with published human kinetic parameters were then used for modeling human exposures of lymphatictargeting Glyph prodrugs. Results illustrate the interspecies scaling of GlyphAllo and support the application of this method to predict PK across species for other Glyph molecules in development.