Extended Evaluation of Virological, Immunological and Pharmacokinetic Endpoints of CELADEN: A Randomized, Placebo-Controlled Trial of Celgosivir in Dengue Fever Patients

Abstract

CELADEN was a randomized placebo-controlled trial of 50 patients with confirmed dengue fever to evaluate the efficacy and safety of celgosivir (A study registered at ClinicalTrials.gov, number NCT01619969). Celgosivir was given as a 400 mg loading dose and 200 mg bid (twice a day) over 5 days. Replication competent virus was measured by plaque assay and compared to reverse transcription quantitative PCR (qPCR) of viral RNA. Pharmacokinetics (PK) correlations with viremia, immunological profiling, next generation sequence (NGS) analysis and hematological data were evaluated as exploratory endpoints here to identify possible signals of pharmacological activity. Viremia by plaque assay strongly correlated with qPCR during the first four days. Immunological profiling demonstrated a qualitative shift in T helper cell profile during the course of infection. NGS analysis did not reveal any prominent signature that could be associated with drug treatment; however the phylogenetic spread of patients' isolates underlines the importance of strain variability that may potentially confound interpretation of dengue drug trials conducted during different outbreaks and in different countries. Celgosivir rapidly converted to castanospermine (Cast) with mean peak and trough concentrations of 5727 ng/mL (30.2 μM) and 430 ng/mL (2.3 μM), respectively and cleared with a half-life of 2.5 (± 0.6) hr. Mean viral log reduction between day 2 and 4 (VLR2-4) was significantly greater in secondary dengue than primary dengue (p = 0.002). VLR2-4 did not correlate with drug AUC but showed a trend of greater response with increasing Cmin. PK modeling identified dosing regimens predicted to achieve 2.4 to 4.5 times higher Cmin. than in the CELADEN trial for only 13% to 33% increase in overall dose. A small, non-statistical trend towards better outcome on platelet nadir and difference between maximum and minimum hematocrit was observed in celgosivir-treated patients with secondary dengue infection. Optimization of the dosing regimen and patient stratification may enhance the ability of a clinical trial to demonstrate celgosivir activity in treating dengue fever based on hematological endpoints. A new clinical trial with a revised dosing regimen is slated to start in 2016 (NCT02569827). Furthermore celgosivir's potential value for treatment of other flaviruses such as Zika virus should be investigated urgently.

Trial registration: ClinicalTrials.gov NCT01619969.

Conflict of interest statement

CS, SW, APSR, KWKC, EEO, JGHL and SGV are named inventors on a patent (Novel Dosing Regimens of Celgosivir For The Treatment Of Dengue; filing date Dec 4, 2013). CS reports personal fees from 60 Degrees Pharmaceuticals for examining pharmacokinetic modelling and dosing regimen for dengue antiviral drugs. JGHL and SGV received research support from 60° Pharmaceuticals PLC and Ministry of Health of Singapore Category 1 grant for extended non-clinical studies on celgosivir.

Figures

Fig 1. Subject disposition for CELADEN study [17].

Fig 2. Correlation between viremia by qPCR and plaque assay.

Red filled circles–patients who received celgosivir; blue open squares–patients who received placebo; lines–linear regression to all data. Pearson correlation coefficient for all data (A)Day 1 0.79 (95% CI: 0.64 to 0.87); (B) Day 2 0.79 (95% CI: 0.65 to 0.87); (C) Day 3 0.83 (95% CI: 0.72 to 0.90); (D) Day 4 (95% CI: 0.55 to 0.83); (E) Day 5 0.36 (95% CI: 0.08 to 0.58).

Fig 3. Viremia kinetics in primary and secondary dengue patients.

Log viremia Mean (± SEM) by day and prior dengue infection status. Green solid circles connected by solid line–primary dengue; open orange squares connected by dashed line–secondary dengue. Inset: VLR 2–4 by prior dengue infection status: box 25th to 75th percentile, whiskers minimum and maximum values. Virus is cleared significantly faster in secondary dengue compared to primary dengue (p = 0.002).

Fig 4. PK profile of castanospermine (semi-log plot).

Solid brown line is the predicted concentration of castanospermine based on the mean PK parameters and the dosing regimen studied in the trial. Gray open circles are the observed peak and trough concentrations of castanospermine. Symbols and error bars are the mean and SEM, respectively. Black dotted line is the target trough concentration (400 ng/mL) predicted based on animal efficacy studies.

Fig 5. Dependence of pharmacokinetic parameters on covariates.

Body Weight (A and B); Age (C and D); Creatinine Clearance (E); and Sex (F). Clearance or volume of distribution were not significantly affected by patients’ body weight, age or sex. Drug clearance was significantly correlated with creatinine clearance, indicating a significant role of the kidneys for elimination of celgosvir. Solid line-linear regression, dashed line- 95% CI. The slope of the linear regression line of creatinine clearance versus drug clearance was 0.86 (95% CI: 0.376, 1.351).

Fig 6. Scatterplots of drug exposure (Cmin, Cmax, AUC) and VLR2-4.

Blue open squares–individual patients who received placebo; red filled circles–individual patients who received celgosivir, separated into two quantiles of exposure. solid heavy line–median value, error bars– 25th to 75th percentile.

Fig 7. Predicted exposure for different dosing regimens.

The Box-25th to 75th percentile, whiskers-minimum and maximum values for the various dosing regimens is shown. (A) Cmin range for the various dosing regimens shows that 150 mg every 6 hr is predicted to yield a 4.5-fold increase in median Cmin used in CELADEN trial (B) Cmax, range do not vary significantly for the various dosing regimens and (C) AUC only shows a modest 1.33-fold increase over the dosing regimen used in the CELADEN trial.

Fig 8. Changes in platelets and hematocrit for celgosivir (red filled circles) or placebo (blue open squares) treated patients.

Mean ± SEM changes in platelets count (A) and hematocrit (B) at different study days in all patients. (C) and (D) are Mean ± SEM changes in platelets count (C) and hematocrit (D) at different study days for secondary dengue patients only. Platelet nadir values (E) and difference between maximum and minimum hematocrit values (F) for secondary dengue patients treated with celgosivir or placebo, solid line–median, bars—interquartile range.

Fig 9. Celgosivir treatment changes the systemic immune response during acute dengue infection.

A) Scatter plot of cytokine concentration of patient plasma samples drawn at specific time points post admission were analyzed for 41 cytokines and chemokines using the Human cytokine panel 1. The standard for each analyte between 4–10000 pg/ml is presented as yellow dashes, quality control for each analyte is in blue filled circles, a healthy volunteer (JEC) sample is shown in black filled circles the patient samples (400 samples/analyte; 8 per patient x 50 patients) are shown as green crosses. B) Heat map of log2 fold changes of averaged analyte concentrations at specific time points against the pre-treatment concentrations for selected analytes showing qualitative shifts in the analyte concentrations. Treatment and control groups fold changes are shown in separate rows for each analyte to aid comparison. Fold changes are colored from over expression in red to under expression in green. A scale showing the value of the maximum and minimum log2 fold change is shown.

Fig 10. Evolutionary relationship of the whole genome of dengue samples isolated from celgosivir treatment and placebo in relation to representative samples collected from Asia.

Maximum likelihood trees were generated using the nucleotide alignment from start to stop codon of the coding region. Tip labels are coloured based on sample type, and genotypes are labelled adjacent to tip labels for DENV 2, and along branches for DENV 1 and 3. Scale bar represents nucleotide substitutions per site.