Identification, preclinical profile, and clinical proof of concept of an orally bioavailable pro‐drug of miridesap

Abstract

Background and Purpose

Miridesap, a depleter of serum amyloid P component (SAP), forms an essential component of a novel approach to remove systemic amyloid deposits; low oral bioavailability necessitates that it is given parenterally. We sought to identify and clinically characterise a pro‐drug that preserves the pharmacological properties of miridesap while having adequate oral bioavailability and physical stability.

Experimental Approach

We utilised a preclinical screening cascade focused on appropriate physicochemical properties, physical and gut stability, and conversion to miridesap in liver microsomes and blood. GSK3039294 (GSK294) had the desired in vitro profile and progressed to preclinical in vivo pharmacokinetic and safety assessments. Based on a favourable profile, it was tested in healthy participants after single and repeat dosing.

Key Results

GSK294 was highly soluble and stable in simulated gastric and intestinal fluids, stable in intestinal microsomes, and permeable in Madine Darby Canine Kidney type II cells. GSK294 was rapidly hydrolysed to miridesap and its mono pro‐drug ester in blood and liver microsomes. GSK294 showed good oral bioavailability of miridesap in rats and dogs. Following administration of GSK294 600 mg QD for 7 days in humans, pharmacodynamically active concentrations of miridesap were achieved with substantial and sustained depletion of plasma SAP. The study was terminated due to observations of arrhythmia, the relation of which to GSK294 remains unclear.

Conclusion and Implications

Using a preclinical screening cascade, we identified a pro‐drug for a palindromic molecule with unique pharmacology (miridesap). The pro‐drug depleted circulating SAP with a time course and extent similar to that of parenterally administered miridesap.

Abbreviations

AUC

area under the concentration time curve

CHI

chromatographic hydrophobicity index

ChromlogD

chromatographic logD

CI

confidence interval

C_max

maximum serum concentration

CVb

between‐participant variability

LLQ

lower limit of quantification

MDCK‐II

Madine Darby Canine Kidney type II

PD

pharmacodynamic

PK

pharmacokinetic

SAP

serum amyloid P component

SGF

simulated gastric fluid

SVT

supraventricular tachycardia

t_max

time at which the maximum concentration is observed

UPLC

ultraperformance LC

What is already known

• Miridesap is a depleter of circulating SAP in humans but has to be given parenterally.

What this study adds

• We describe preclinical identification of an orally bioavailable pro‐drug of miridesap and characterisation in humans.

What is the clinical significance

• GSK294 was comparable to parenteral miridesap regimens used in clinical studies in systemic amyloidosis.

1. INTRODUCTION

Serum amyloid P component (SAP) is a non‐fibrillar, plasma glycoprotein that circulates at a concentration of 20–40 mg·L⁻¹ and is a normal constituent of the extracellular matrix located on the microfibrillar mantle of elastic fibres throughout the body and in the lamina rara interna of the glomerular basement membrane. In systemic amyloidosis, a clinical disorder that is caused by a progressive accumulation of insoluble, misfolded deposits of normally soluble precursor proteins (amyloid) in the extracellular matrix of target organs, SAP universally decorates amyloid deposits and contributes to their persistence. Professor Pepys identified SAP as a drug target in systemic amyloidosis, which led to the development of miridesap (Pepys et al., 2002) and subsequently dezamizumab (anti‐SAP monoclonal antibody) for clinical testing (Bodin et al., 2010; Richards et al., 2015; Richards et al., 2018). Miridesap is a small molecule depleter of circulating SAP with a unique mechanism of action; it non‐covalently crosslinks pentameric SAP molecules to form a decameric complex that is rapidly removed from the circulation (Pepys et al., 2002). In clinical studies to date, it has been well tolerated with predictable pharmacokinetic/pharmacodynamic (PK/PD) characteristics (Gillmore et al., 2010; Kolstoe & Wood, 2010; Richards et al., 2015; Richards et al., 2018; Sahota et al., 2015). Despite these favourable characteristics, its clinical utility may be limited by the fact that it must be given parenterally, owing to very low oral bioavailability. Miridesap is a hydrophilic diacid, and we hypothesised that it was the presence of the two acid moieties that were not only essential to binding to SAP but also responsible for severely restricted absorption following oral dosing. We therefore sought to identify a pro‐drug with good oral bioavailability and, upon entering the circulation, converts quickly and completely to miridesap, thereby preserving the favourable PK/PD characteristics of miridesap. Initial attempts to generate a potentially orally bioavailable double acid pro‐drug of miridesap (Huwyler, Jakob‐Roetne, & Poli, 2003) did not lead to a drug.

Here, we describe a successful screening cascade approach that identified a pro‐drug molecule, GSK3039294 (GSK294) with the desired balance of physical properties and stability, while preserving the unique pharmacology of miridesap. Unexpected safety observations in the clinical study, although not thought to be related to the pharmacology of miridesap, limited preliminary evaluation in humans.

2. METHODS

2.1. Materials

The compound GSK294 was synthesised and analysed according to the published method (Denis & Mirguet, 2015).

Control‐naïve blood and plasma were ethically sourced from approved vendors or within GlaxoSmithKline (GSK); all in vivo studies were ethically reviewed according to local site practices under specific animal protocols in AAALAC accredited facilities in accordance with GSK's policies on the Care, Welfare and Treatment of laboratory animals in the UK; the Animals (Scientific Procedures) Act, 1986; and the Human Tissue Act, 2004. Individual in vitro and preclinical in vivo experiments were conducted with fewer than n = 5 animals in accordance with the 3Rs (replacement, reduction, and refinement). Han‐Wistar male rats (10–12 weeks old with body weight of 200–250 g) and Marshal beagle male dogs (at least 10 months old with body weight of 6–15 kg) were supplied by Charles River UK or France, or Harlan UK. These animals were selected as the intended toxicological species as required by most regulatory guidelines to provide safety and systemic exposure information using the intended clinical route of oral administration prior to dosing in the clinic. All animals were housed in standard conditions and provided with food and water ad libitum. Animal studies are reported in compliance with the ARRIVE guidelines (Kilkenny, Browne, Cuthill, Emerson, & Altman, 2010) and with the recommendations made by the British Journal of Pharmacology.

The Madine Darby Canine Kidney type II (MDCK‐II) cells were obtained from America Type Culture Collection (MDCK.2 [ATCC® CRL2936™]).

2.2. Data and statistical analysis

The data and statistical analysis comply with the recommendations of the British Journal of Pharmacology on experimental design and analysis in pharmacology (Curtis et al., 2018). Statistical analyses were not conducted on the preclinical results of this study due to low sample size (n < 5). PK parameters were determined by non‐compartmental analyses using Phoenix WinNonlin Enterprise software (Certara, Princeton, NJ).

2.3. Analytical methods

2.3.1. Chromatographic logD determination

The chromatographic hydrophobicity index (CHI; Valko, Bevan, & Reynolds, 1997) values were measured using reversed phase HPLC column (50 × 2 mm, 3‐μM Gemini NX C18, Phenomenex, UK) with fast acetonitrile gradient at starting mobile phase of pHs 2, 7.4, and 10.5. CHI values are derived directly from the gradient retention times by using a calibration line obtained for standard compounds. The CHI value approximates to the volume % organic concentration when the compound elutes. CHI is linearly transformed into chromatographic logD (ChromlogD; Young, Green, Luscombe, & Hill, 2011).

2.3.2. Solubility in physiological solutions

Compound sample was accurately prepared to approximately 10 mg·ml⁻¹ in the relevant simulated physiological fluid, prepared according to established protocols (Butler & Dressman, 2010) and mechanically agitated at ambient temperature (21–23°C) for 24 hr. Samples were taken at 0.5, 4, and 24 hr, solid removed by filtration, and the solution quantitatively analysed using ultraperformance LC (UPLC), with a standard solution of the analyte as reference.

2.3.3. Physical stability in solution

The assay was designed to determine physical stability of a compound in buffer at various pHs. GSK294 was dissolved in DMSO at 1 mg·ml⁻¹. PBS was prepared by mixing K₂HPO₄ 50 mM and KH₂PO₄ 50 mM solutions to obtain four buffers at pHs 6.0, 7.0, 7.5, and 8.0.

Stability studies were conducted at room temperature (25°C) and were started by the addition of 8 μl of DMSO solution to 792 μl of PBS 50 mM at pH 6.0, 7.0, 7.5, or 8.0 (1% DMSO) to give 10 μg·ml⁻¹ final solutions; 75 μl of mixture was taken at 0, 1, 2, 4, and 24 hr, and 225‐μl acetonitrile containing internal standard was added prior to LC–MS/MS; conditions are described below.

2.3.4. Blood hydrolysis in rat, dog, and human

The assay was designed to determine stability of compound in fresh blood. Compound was dissolved in DMSO at 1 mg·ml⁻¹. Daughter solution was prepared in DMSO at 100 μg·ml⁻¹. Fresh blood was diluted by ½ in isotonic physiological pH 7.4 buffer (n = 2).

Pre‐incubation consisted of warming 495 μl of blood at 37°C for 7 min. Incubations were started with the addition of 5 μl of daughter solution; 50 μl of mixture were taken at 0, 5, 15, 30, and 60 min; 50 μl of water was added to sample and then quenched with 300‐μl acetonitrile containing internal standard and centrifuged (10 min at 4,000 r.p.m.) prior to analysis by LC–MS/MS.

2.3.5. Intestinal and liver microsomal hydrolysis assay

The assay was designed to determine stability of compound in intestinal and liver microsomal matrices (n = 2). Compound was dissolved in DMSO at 1 mg·ml⁻¹. Daughter solution was prepared in methanol/water (50/50) at 30.3 ng·ml⁻¹. Human, rat, and dog and liver and intestinal microsomes (from Xenotech) were diluted at 0.625 mg protein·ml⁻¹ in PBS 50 mM pH 7.4 (0.5 mg·ml⁻¹).

Pre‐incubation consisted of warming microsomal solution 395 μl with 100 μl of NADPH in NaHCO₃ (2%) at 37°C for 7 min. Incubation was started with the addition of 5 μl of daughter solution; 50‐μl aliquots of mixture were taken at 0, 3, 6, 12, and 30 min and quenched with 150‐μl acetonitrile containing internal standard. After 10 min centrifugation at 4,000 r.p.m., 2 μl of samples were analysed by LC–MS/MS.

2.3.6. Passive diffusion permeability of GSK294 in MDCK‐II cells at pH = 7.4

An assessment of intestinal absorption was conducted using the MDCK‐II cell line (ATCC), which were cultured and amplified to determine the passive permeability of GSK294. This was based on the published validated methodology (Thiel‐Demby et al., 2009) where basolateral to apical (B → A) directional transfer of GSK294 (3 μM) was tested in triplicate and incubated for 90 min at pH 7.4. Lucifer yellow was used as a paracellular marker to determine the integrity of MDCK‐II monolayer and propranolol and atenolol used as positive controls. All samples were diluted 1:1 with acetonitrile, and test drug concentration was determined using LC–MS/MS analysis.

2.4. Preliminary pharmacokinetics

2.4.1. Intravenous administration of GSK294 in male Han‐Wistar rats and male beagle dogs

GSK294 in solution in DMSO/hydroxypropyl‐β‐cyclodextrin 20% in PBS 60 mM pH 7 (5:95) was administered as a single intravenous bolus injection to non‐fasted rats (n = 2 at 10 mg·kg⁻¹, 2 ml·kg⁻¹) and to dogs (n = 1 at 1 mg·kg⁻¹, 1 ml·kg⁻¹) with food given 2 hr post‐administration.

Serial blood samples were collected via tail vein (rat) or jugular vein (dog) at intervals up to 24 hr post‐administration, respectively. Aliquots of whole blood were diluted 50:50 with water and stored at approximately −20°C prior to LC–MS/MS analysis using a method based upon protein precipitation with acetonitrile and a generic internal standard.

2.4.2. Oral bioavailability of GSK294 in male Han‐Wistar rats and male beagle dogs

GSK294A in HPMC K100 0.5%/Tween 80 0.1% in PBS 60 mM pH 7 was administered by oral gavage at doses ranging from 3 to 100 mg·kg⁻¹ (n = 2 at 10 ml·kg⁻¹) to non‐fasted rats and to dogs at (5 mg·kg⁻¹, 5 ml·kg⁻¹) with food given 2 hr post‐administration with blood sampled as above across a 24‐hr time period and stored at −20°C prior to LC–MS/MS analysis.

2.5. Toxicokinetics

2.5.1. Oral assessment of GSK294 in male Han‐Wistar rats and male beagle dogs

Using the same oral formulation, GSK294 was administered by oral gavage as part of the toxicological safety assessment at doses ranging from 100 to 1,000 mg·kg⁻¹ (n = 3 at 10 ml·kg⁻¹) to non‐fasted rats for up to 6 months and to dogs up to 28 days (n = 3 at 5 ml·kg⁻¹) with food given 2 hr post‐administration with blood sampled across a 24‐hr period on each sampling occasion.

Serial blood samples were collected into EDTA tubes containing BNPP chilled on ice, and triplicate aliquots were precipitated immediately with acetonitrile containing internal isotopic stable standard for GSK294, with the extracted solvent removed from the blood pellet and stored at approximately −20°C prior to LC–MS/MS analysis. Each remaining blood sample was also immediately centrifuged (10 min at 4,000 g, at +4°C) to yield plasma and aliquots precipitated with internal isotopic stable standard for miridesap and stored at −20°C prior to LC–MS/MS analysis.

2.6. Clinical methods

This single centre (GSK Clinical Unit), open‐label, non‐randomised, first‐in‐human study of the GSK294 pro‐drug was approved by the UK Medicines and Healthcare Products Regulatory Agency and the Berkshire “B” phase 1 accredited Research Ethics Committee. Part A was a single dose escalation over three dose levels; the two cohorts received two dose levels (200 and 600 mg; or 600 and 1,200 mg, respectively). Part B consisted of 7 days repeat dosing (600 mg OD) in a single cohort. Sentinel dosing was employed in both study parts.

Only healthy participants aged 18–70 years who provided informed consent were eligible for the trial. Health status was assessed at screening based on medical history, physical examination, cardiac monitoring, and laboratory testing. Participants were monitored in house for 3 days in Part A and 7 days in Part B. Follow‐up was conducted 7–14 days post‐dose. Detailed safety assessments consisting of adverse event monitoring, ECGs, cardiac telemetry, vital signs, urinalysis, haematology, and biochemistry were undertaken.

Blood and plasma samples for PK analysis (GSK294 and miridesap, respectively) were taken at the following time points in Part A: pre‐dose, 0.25, 0.5, 0.75, 1, 2, 3, 4, 5, 6, 8, 10, 12, 16, 24, and 48 hr and processed as described for preclinical toxicokinetic studies. Samples for measurement of plasma SAP were taken in Part B only at pre‐dose and 2 hr on Day 1, pre‐dose on Days 2 and 4, pre‐dose, 0.5, 1, 2, 3, 4, and 6 hr post‐dose on Day 5, and again pre‐dose on Day 7 and off‐dose on Day 14.

Further repeat dose cohorts were planned and a Part C (repeat dosing in patients with amyloidosis), but the study was terminated before these were conducted owing to arrhythmia findings that require further evaluation.

2.7. Sample analysis

Samples were analysed by LC–MS/MS with an electrospray source (API5000), in positive mode and with following mass transitions: GSK294: 657 to 384; monoester of miridesap: 499 to 226; and miridesap: 341 to 226 or 198.

For early in vitro (physical, physiological, blood, and microsomal stability) samples and preliminary in vivo PK (rat and dog) sample extracts were injected on an Acquity UPLC system (Waters) and eluted on an Ascentis Express C18 column (50 × 2.1 mm id, 2.7 μm) at 0.5 ml·min⁻¹ at 50°C, with 0.1% formic acid in water (A) and 0.1% formic acid in acetonitrile (B), using the following elution 2 min gradient: 5% to 95% B over 1.5 min, 95% B over 0.5 min, and 0.1 min for re‐equilibrate column. Controls were made to calculate percentage of disappearance of parent but also appearance of suspected metabolite (i.e., monoester and diacidic form against appropriate calibration lines).

Toxicokinetic (rat and dog) and human clinical samples were quantitatively analysed using a fully validated assay using a 50‐μl aliquot of sample. The lower limit of quantitation was 10 ng·ml⁻¹, and the upper limit of quantitation was 5,000 ng·ml⁻¹ in blood for GSK294 extracted using 0.1% formic acid in acetonitrile and in plasma for miridesap extracted using 0.1% formic acid in methanol containing appropriate internal standard. Samples were injected on an Acquity UPLC system (Waters) and eluted on an BEH Phenyl C18 column (50 × 2.1 mm id, 1.7 μm) at 0.6 ml·min⁻¹ at 50°C, with 0.1% formic acid in water (A) and acetonitrile (B), using the following elution 2 min gradient: held at 5% B for 0.2 min and then 5% to 40% B for 0.8 min and increased to 90% B by 1.3 min and held for 0.4 min, before re‐equilibrating column for 0.2 min.

Non‐compartmental methods were used for PK analysis of blood or plasma concentration versus time data using Phoenix WinNonlin Enterprise (Certara).

For total SAP biomarker concentration, clinical study samples from participants dosed with GSK294 were quantitatively analysed in a fully validated electrochemiluminescent immunoassay method using the MSD QuickPlex platform and software Discovery Workbench (v4.0). Plasma samples were diluted 1:1,000 and a further 1:2‐fold dilution when mixed with assay buffer and antibody cocktail (biotinylated capture and sulfo‐tag detection), incubated, and transferred to Streptavidin Gold MSD plates. Using a 70‐μl diluted aliquot of sample, the lower limit of quantitation was 0.065 ng·ml⁻¹, and the upper limit of quantitation was 100 ng·ml⁻¹ using SAP depleted plasma for calibration curve preparation. Therefore, to obtain the actual SAP concentration in each sample, a back‐calculated multiplier was applied.

2.8. Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Harding et al., 2018), and are permanently archived in the Concise Guide to PHARMACOLOGY 2015/16 (Alexander et al., 2019).

3. RESULTS

3.1. GSK294 preclinical profile

3.1.1. In vitro lipophilicity, solution stability, solubility, and permeability

Physicochemical properties for GSK294 are shown in Table 1, indicating a log D in the desired range for oral bioavailability (Lipinski, Lombardo, Dominy, & Feeney, 1997), high solubility in simulated gastric and intestinal fluids, and high permeability through MDCK‐II cell monolayer (Thiel‐Demby et al., 2009) commensurate with a Developability Classification System class I molecule, the class most likely to be developable as an orally administered drug based on precedent with other molecules (Butler & Dressman, 2010; Rosenberger, Butler, & Dressman, 2018).

Table 1.

Physicochemical and physiological properties of GSK294

Lipophilicity ChromlogD	Solubility (mg·ml−1)			MDCK‐II passive permeability (nm·s−1)
	SGF (pH 1.6)	FaSSIF	FeSSIF	Papp (nm·s−1)	Pexact (nm·s−1)
3.6	5.2, 5.0	4.4, 4.8	5.0, 3.8	133	148

Abbreviations: ChromlogD, chromatographic logD; FaSSIF, fasted state simulated intestinal fluid; FeSSIF, fed state simulated intestinal fluid; MDCK‐II, Madine Darby Canine Kidney type II cells; P_app/exact, apparent/exact permeability; SGF, simulated gastric fluid.

The solution stability of GSK294 decreases notably at pH 8, with the main route of degradation via hydrolysis to the monoester (intermediate 1) and miridesap (Figure 1). However, in simulated gastric fluid (SGF; pH 1.6, data not shown) and at pH 6 (the pH expected in the intestine), GSK294 has a long half‐life of degradation (145–179 hr).

Figure 1.

In vitro physical hydrolysis profile of GSK294 at varying pH in aqueous solution

3.1.2. Hydrolysis in blood and in intestinal and liver microsomes

To gain insight to the potential for stability to intestinal enzymes during absorption and then cleavage of the pro‐drug to miridesap once absorbed, the hydrolysis of GSK294 was evaluated in rat, dog, and human intestinal and liver microsomes and blood (Figure 2). Hydrolysis rates in intestinal microsomes are low, with low intrinsic clearance rates, suggesting that the pro‐drug might not be cleaved by this route during absorption. Conversely, in liver microsomes, substantial cleavage of the pro‐drug GSK294 to miridesap via the monoester intermediate 1 is predicted in all species, with high intrinsic clearance rates, suggesting that GSK294 might be a suitable pro‐drug for cleavage to miridesap once absorbed. The rate of hydrolysis was also high in rat blood, suggesting a potential additional systemic route of cleavage of pro‐drug, although hydrolysis was lower in dog blood and in human blood.

Figure 2.

In vitro matrix hydrolysis profile of GSK294 in intestinal and liver microsomes and blood

These data suggested that GSK294 would be absorbed in vivo following administration by the oral route and then cleaved to afford miridesap once systemically available. The PK profile of the pro‐drug was therefore evaluated in vivo.

3.1.3. Preliminary pharmacokinetics in rats and dogs

Miridesap generated by oral administration of GSK294 has a high equivalent oral bioavailability in the rat (56–80% at 3 mg·kg⁻¹ and 36–58% at 10 mg·kg⁻¹) although only 22% equivalent bioavailability in dog at a similar dose level (5 mg·kg⁻¹; Table S1). At these doses, the pro‐drug GSK294 was not detected up to 24 hr post‐oral administration in the two preclinical species, indicating a rapid in vivo cleavage once absorbed. The monoester intermediate 1 was detected up to 2 hr post‐oral administration only in the dog but at a very low concentration.

Additional preliminary PK studies in rats at 10, 30, and 100 mg·kg⁻¹ (n = 2) indicated similar release of miridesap with bioavailability of 50 ± 21%, 37 ± 8%, and 30 ± 6%, respectively. At doses up to 100 mg·kg⁻¹ in the rat, GSK294 showed a roughly dose‐proportional increase in miridesap exposure and maximum plasma concentration (C_max). Monoester intermediate 1 was detected only at the top dose and up to 2–4 hr post‐oral administration.

Toxicokinetic repeat dose studies (100, 300, and 1,000 mg·kg⁻¹; n = 3) further supported consistent high exposure of miridesap in both rat and dog but in a less than dose‐proportional manner, indicating absorption as a potential rate‐limiting factor at such doses (Figure 3 for rat and Figure 4 for dog). GSK294 and monoester intermediate were not measurable in rodent blood due to the rapid systemic hydrolysis. In the dog, GSK294 was quantifiable, and exposure increased with increasing dose, albeit with an “intravenous‐like” profile that rapidly declined with time and with very low levels detectable up to 24 hr (data not shown).

Figure 3.

Plasma pharmacokinetic profile of miridesap after intravenous administration of miridesap at 10 mg·kg⁻¹ and after oral administration of GSK294 at a range of doses in rat. LLQ, lower limit of quantification

Figure 4.

Plasma pharmacokinetic profile of miridesap after intravenous administration of miridesap at 10 mg·kg⁻¹ and after oral administration of GSK294 at a range of doses in dog. LLQ, lower limit of quantification

Thus, GSK294 is soluble, permeable, and stable across species in intestinal microsomes, sufficient to enable oral delivery and absorption to the systemic circulation. It is hydrolysed rapidly in vitro across species in blood and in liver microsomes, and this is reflected in a good miridesap PK profile in both rat and dog following oral dosing of GSK294. These data suggest that the clinical PK profile would also be appropriate.

3.2. Clinical study

All recruited participants were male, 17 in Part A (mean age 41 years) and 8 in Part B (mean age 35 years). Most participants were White; one was Black, and one Asian. One participant was withdrawn owing to an adverse event, all others completed the protocol as planned. The 50‐year‐old participant who was withdrawn experienced an atrial tachycardia 1 day after administration of a single dose of 600 mg of GSK294 and further experienced an accelerated idioventricular rhythm on the same day. Both were self‐limiting and not associated with cardiovascular compromise. One further participant in Part A experienced a self‐limiting period of atrial tachycardia following administration of 1,200 mg of GSK294. In response to these observations, cardiac monitoring was increased in Part B.

One serious AE was reported in Part B. The 39‐year‐old male participant experienced short, recurring episodes of supraventricular tachycardia (SVT, accelerated junctional rhythm/junctional tachycardia), which became mildly symptomatic from Day 6 onwards and resolved on Day 10. Episodes occurred at rest with symptoms of mild palpitations. All episodes were brief, self‐terminating, and not associated with cardiovascular compromise. The participant's drug exposure and degree of SAP depletion were consistent with other participants. No clinically significant changes in ECG were reported.

The study was temporarily halted at this point. Subsequent review concluded that the design of the remainder of the study did not lend itself to adequate evaluation of the cardiac profile of GSK294 and therefore the study was formally terminated. No other pattern of adverse events or trends in clinical laboratory parameters suggestive of a safety signal was observed in this limited study (Table S2).

3.2.1. Clinical pharmacokinetics and pharmacodynamics

Following dosing with GSK294 at single doses of 200, 600, or 1,200 mg, GSK294 concentrations were mostly non‐quantifiable or very low. At all dose levels, in about half of the participants, it was measurable in one or two time points per participant up to 2 hr post‐administration with maximum GSK294 concentrations of about 2–3× the lower limit of quantification of the analytical assay (LLQ 10 ng·ml⁻¹). All other participants had concentrations lower than the LLQ at all time points.

As shown in Figure 5, miridesap itself, however, was detectable with maximum concentrations similar to reported concentrations following parenteral administration (Gillmore et al., 2010; Pepys et al., 2002; Sahota et al., 2015), which suggests a good bioavailability of GSK294, followed by a quick and complete conversion into miridesap.

Figure 5.

Miridesap concentration versus time following oral administration of GSK294. LLQ, lower limit of quantification

The PK parameters of miridesap following single oral doses of GSK294 are shown in Table S3. Between‐participant variability (% CVb) in PK parameters was generally low (<30%). Increases in miridesap area under the concentration time curve (AUC) were dose proportional at 200–600 mg of GSK294 and slightly less than proportional at 1,200 mg. C_max displayed slightly less than dose‐proportional increases over the full dose range tested. Terminal half‐lives did not show any trend with dose.

Based on the single dose PK data, a dose of 600 mg·day⁻¹ was selected for the repeat dose part (Part B) of the study, in which plasma SAP was measured. A preliminary assessment of the effect of food was investigated by comparing miridesap concentrations on dosing Day 4 without food, following a standardised (but not high fat) breakfast on dosing Day 5.

GSK294 600 mg·day⁻¹ for 7 days resulted in a rapid depletion of SAP in plasma with reductions of ~85% as early as 2 hr post‐dose. Plasma SAP levels lower than 3 mg·L⁻¹ were achieved from 24 hr post‐dose onwards (see Figure 6). Although a defined target degree of SAP depletion has not been formally established, these concentrations allowed administration of anti‐SAP antibodies in clinical studies in systemic amyloidosis (Richards et al., 2015; Richards et al., 2018).

Figure 6.

Plasma SAP levels after oral administration of GSK294 600 mg·day⁻¹ for 7 days. SAP, serum amyloid P component

PK parameters of miridesap following repeat dosing with pro‐miridesap at 600 mg·day⁻¹ are shown in Table S4. The median t _max (time at which the maximum concentration is observed) of miridesap following oral dosing with GSK294 is somewhat later in the fed state compared with the fasted state, although the ranges completely overlap. The effect of food on systemic exposure of miridesap (C_max and AUC_(0‐tau)) was modest; the increase was 1.5‐ and 1.2‐fold higher respectively in the fed versus fasted state. As in Part A of the study, the majority of samples had concentrations of GSK294 that were below the LLQ.

4. DISCUSSION AND CONCLUSIONS

The pharmacology of miridesap is unique and has the rare distinction of being pharmacologically active in all human subjects in whom it has been tested (Gillmore et al., 2010; Kolstoe & Wood, 2010; Pepys et al., 2002; Richards et al., 2015; Richards et al., 2018; Sahota et al., 2015). Its physicochemical properties limit administration to parenteral routes. We sought to identify a pro‐drug that rapidly and efficiently generated miridesap in the blood following oral administration. Miridesap is a small molecule containing two N‐linked R‐carboxypyrrolidine moieties (Figure 7), which use functional groups with high atom efficiency in binding to SAP such that two molecules of SAP are crosslinked for hepatic clearance. The two carboxylic acid functional groups are key to the binding affinity but are also likely to be responsible for poor membrane permeability and thus oral bioavailability. Analysis of the architecture of the miridesap binding site on SAP indicates that changing the carboxylic acids for more membrane permeable groups while retaining SAP binding affinity would be extremely challenging; an appropriate screening cascade (Figure 8) to balance properties that confer oral bioavailability and physical stability was deemed more likely to succeed.

Figure 7.

Sequential hydrolysis of GSK294 pro‐drug via its monoester derivative to generate miridesap and two molecules each of formaldehyde, carbon dioxide, and tetrahydropyran‐4‐ol (THP‐OH)

Figure 8.

Cascade of assays used to identify GSK294

We explored several pro‐drug functionalities, including esters, amides, ketal esters, and acetal esters, anticipating that, once absorbed, miridesap would be liberated from the pro‐drug through sequential cleavage of the derivatised acid groups by blood and/or liver enzymes. The head groups for these functional groups were selected to manage overall physicochemical properties; ChromlogD was used to measure lipophilicity to benchmark in silico predictors for the chemical series. This enabled chemistry to focus only on molecules, such as GSK294 (ChromlogD 3.6), in drug‐like lipophilicity space (Lipinski et al., 1997). Molecules were screened for solubility and permeability, properties key to efficient and reproducible oral absorption. To limit future developability challenges, we only evaluated molecules that proved to be crystalline, defined qualitatively initially, and subsequently confirmed by X‐ray powder diffraction analysis (data not shown).

Having identified molecules that were crystalline and more likely to be Developability Classification System class 1 (kinetic solubility >100 μg·ml⁻¹; permeability >100 nm·s⁻¹), we tested for solution stability, including at acid pH and in SGF. GSK294 was amongst those compounds exhibiting minimal hydrolysis in acid solution, including SGF. Similarly, we sought to identify compounds with minimal hydrolysis by intestinal enzymes prior to absorption and so screened compounds in rat, dog, and human intestinal microsomes. GSK294 was substantially intact after 1‐hr exposure to intestinal microsomes across species.

Once absorbed, hydrolysis at each ester of the GSK294 molecule is required to liberate miridesap (Figure 4). Esterases are prevalent in blood, intestinal wall, and liver. We selected compounds based on their in vitro profile in rat, dog, and human liver microsomes and in rat, dog, and human blood. GSK294 showed substantial conversion to miridesap in liver microsomes from all species, although conversion in dog liver microsomes was greater and almost complete by 1 hr. in vitro conversion of GSK294 to miridesap was less extensive in blood than in liver microsomes. Other compounds showed greater conversion to miridesap in blood in vitro but also showed greater conversion in the intestinal microsome assay. We focused on compounds that were predicted by our assays to be more stable prior to systemic exposure to mitigate against developability issues such as instability during manufacture and storage. The principal concern at this stage was that monohydrolysis in vivo would liberate a pharmacologically inactive molecule, as miridesap depends on non‐covalent crosslinking for its action (Pepys et al., 2002). We therefore profiled GSK294 in rat and dog in vivo. The results showed rapid and substantial exposure to miridesap following a single dose, although equivalent oral bioavailability of miridesap was higher in the rat than in the dog. Exposure to pro‐drug or the monoester intermediate could not be measured owing to the rapidity of the conversion. We recognised this would limit the ability of the preclinical toxicology studies to assess toxicity of parent and intermediates if the conversion in man proved markedly slower. The estimated bioavailability was variable, but estimates in both species were potentially acceptable if observed in man.

Given the known pharmacology and favourable safety profile of miridesap (Sahota et al., 2015), the initial (single) doses to be tested in man were selected to give a robust PK assessment. Assuming fast and complete oral absorption of GSK294 and fast and complete subsequent conversion of GSK294 into miridesap, both processes can be described by a single first‐order process (Ka) with high bioavailability (F). For the purposes of selecting the starting dose, Ka and F were assumed to be high: 1.3 hr⁻¹ and 100%. Distribution and elimination of miridesap were expected to be the same, whether parenteral miridesap or oral GSK294 administration. Using this model, the concentration time profiles of miridesap after a single dose of GSK294 were simulated, and the predicted C_max and AUC_(0‐inf) were derived. For the starting dose of 200 mg, the predicted AUC_(0‐inf) was 14.5 μg·h mL⁻¹ (95% confidence interval [CI] 9.5, 22.4 μg·h mL⁻¹) and C_max 3.4 μg·ml⁻¹ (95% CI 2.1, 5.0 μg·h mL⁻¹). The predicted exposure was well within the exposure limit of AUC_(0–24) of 250 μg·h mL⁻¹ and C_max of 22.1 μg·ml⁻¹ set by the no observed adverse effect level in non‐clinical toxicology studies. Observed AUC and C_max in man, at a dose of 200 mg, were 8.0 μg·h mL⁻¹ (95% CI 6.3, 10.0 μg·h mL⁻¹) and 0.72 μg·ml⁻¹ (95% CI 0.59, 0.87 μg·h mL⁻¹), respectively. Consistent with the preclinical profile, exposure to GSK294 was low, and a formal exposure profile was not observed with any administered dose, despite an LLQ of 10 ng·ml⁻¹. The preclinical profile of GSK294 was predictive of that seen in human, with the profile in rat more representative of the human profile than that seen in dog.

Repeat dosing of GSK294 600 mg·day⁻¹ resulted in a rapid and sustained reduction in plasma SAP, with reductions of ~85% achieved as soon as 2 hr post‐dose. In studies with an anti‐SAP monoclonal antibody, intravenous miridesap was used to reduce plasma SAP to 3 mg·L⁻¹ or lower before administering the monoclonal antibody. Following repeat dosing of GSK294, this target was met at 24 hr and maintained for the dosing period of 7 days.

The hydrolysis of GSK294 results in the liberation of formaldehyde. Formaldehyde is an in vitro mutagen but also a by‐product of many biological reactions, and humans are routinely exposed to formaldehyde from the environment, food, and as an endogenous metabolite. There are several marketed pro‐drugs (including fosphenytoin, tenofovir disoproxil fumarate, and fospropofol disodium) that generate formaldehyde in vivo without evidence of harm to human health. As a precaution, we limited the dose to 1,200 mg daily, which would be predicted to liberate 110‐mg formaldehyde, an amount similar to fospropofol disodium. The other molecule generated by cleavage of the pro‐drug is 4‐hydroxy‐tetrahydropyran (Figure 7), which is not well characterised but did not have structural features that indicated specific risks.

In Part A of the clinical study, a total of four separate arrhythmia events were reported in three participants aged 46–55 years: three episodes of SVT (at 600 and 1,200 mg), all consistent with an atrial tachycardia morphology and one episode of a single 5‐beat accelerated idioventricular rhythm (600 mg). Each event was of short duration, asymptomatic, and not associated with haemodynamic instability. SVT is more common in those aged >45 years compared with those <45 years (13.9% vs. 1.1%, respectively; Hingorani, Karnad, Rohekar, Lokhandwala, & Kothari, 2016). There was no association between the time of onset and concentration of GSK294 or miridesap. As a precaution, cardiac monitoring was increased in Part B. In Part B, one participant experienced junctional tachycardia necessitating prolonged inpatient monitoring and thus fulfilled the definition of a serious adverse event. Again, there was no relationship between the onset of the arrhythmia and peak plasma concentration of GSK294 or miridesap. Follow‐up exercise testing, echocardiogram, and 48‐hr Holter recording were normal. Miridesap has not been associated with arrhythmias in clinical studies to date, including patients with cardiac amyloidosis (Gillmore et al., 2010; Richards et al., 2015; Richards et al., 2018). Preclinical profiling against a wide range of ion channels and in vivo pharmacology studies did not identify a pro‐arrhythmic potential for GSK294 or miridesap. The overall assessment of these events was that, while GSK294 was unlikely to be responsible, these unusual findings warranted further evaluation. The optimal design for this would be a placebo‐controlled crossover, and the current study design did not lend itself to this approach. On that basis, the study was terminated early.

Using a preclinical screening cascade, we identified a pro‐drug that showed good oral bioavailability of miridesap. Preclinical results proved predictive of the profile in humans. GSK294 had a favourable PK/PD profile in preliminary human testing indicating that it could be a viable replacement for parenteral miridesap. A higher than expected incidence of supraventricular arrhythmias, although not clearly related to GSK294, curtailed the clinical study and requires further evaluation in future clinical studies.

CONFLICT OF INTEREST

M.B., L.L., N.H., G.L., J.S., D.F., S.K., D.T., and D.H. are employees of GSK and hold stocks/shares. D.R. was an employee of GSK at the time of the study.

AUTHOR CONTRIBUTIONS

M.B., J.S., D.R., D.S.H., N.H., D.T., G.L., and L.L. contributed to the conception and design of the study. D.F. and S.K. contributed to the acquisition of data. M.B., J.S., D.R., D.S.H., N.H., D.T., G.L., L.L., and S.K. contributed to data analysis and/or interpretation. All authors were involved in development of the manuscript and approved the final version.

DATA SHARING STATEMENT

Anonymised individual participant data and study documents can be requested for further research from http://www.clinicalstudydatarequest.com.

DECLARATION OF TRANSPARENCY AND SCIENTIFIC RIGOUR

This Declaration acknowledges that this paper adheres to the principles for transparent reporting and scientific rigour of preclinical research as stated in the BJP guidelines for https://bpspubs.onlinelibrary.wiley.com/doi/full/10.1111/bph.14207, and https://bpspubs.onlinelibrary.wiley.com/doi/full/10.1111/bph.14206, and as recommended by funding agencies, publishers, and other organisations engaged with supporting research.

Supporting information

Table S1. PK parameters of miridesap after the oral administration of GSK294 in preclinical species

Table S2. Summary of all AEs by preferred term

Table S3. Miridesap PK parameters following oral administration of GSK294

Table S4. Miridesap PK parameters following repeat oral administration of pro‐miridesap 600 mg day⁻¹

Richards D, Bamford M, Liefaard L, et al. Identification, preclinical profile, and clinical proof of concept of an orally bioavailable pro‐drug of miridesap. Br J Pharmacol. 2020;177:1853–1864. 10.1111/bph.14956