Abstract
Sparsentan is a single‐molecule dual endothelin angiotensin receptor antagonist (DEARA) currently under investigation as a treatment for focal segmental glomerulosclerosis (FSGS) and IgA nephropathy (IgAN). A population pharmacokinetic (PK) analysis was performed to characterize the PKs of sparsentan and to evaluate the impact of FSGS disease characteristics and co‐medications as covariates on sparsentan PKs. Blood samples were collected from 236 healthy volunteers, 16 subjects with hepatic impairment, and 194 primary and genetic FSGS patients enrolled in nine studies ranging from phase I to phase III. Sparsentan plasma concentrations were determined using validated liquid chromatography–tandem mass spectrometry with a lower limit of quantitation of 2 ng/mL. Modeling was conducted with the first‐order conditional estimation with η–ϵ interaction (FOCE‐1) method in NONMEM. A total of 20 covariates were tested using a univariate forward addition and stepwise backward elimination analysis with significance level of p < 0.01 and p < 0.001, respectively. A two‐compartment model with first‐order absorption and an absorption lag time with proportional plus additive residual error (2 ng/mL) described sparsentan PKs. A 32% increase of clearance due to CYP3A auto‐induction occurred at steady‐state. Covariates retained in the final model included formulation, cytochrome P450 (CYP) 3A4 inhibitor co‐administration, sex, race, creatinine clearance, and serum alkaline phosphatase. Moderate and strong CYP3A4 inhibitors comedications increased area under the concentration‐time curve by 31.4% and 191.3%, respectively. This population PK model of sparsentan suggests that dose adjustments may be warranted for patients taking moderate and strong CYP3A4 inhibitors concomitantly, but other covariates analyzed may not require dose adjustments.
Study Highlights.
WHAT IS THE CURRENT KNOWLEDGE ON THE TOPIC?
Sparsentan is a single‐molecule dual endothelin angiotensin receptor antagonist (DEARA) under investigation as a treatment for focal segmental glomerulosclerosis (FSGS). Patients treated with sparsentan reach an FSGS partial remission end point at higher rates than with the current standard‐of‐care treatment, irbesartan.
WHAT QUESTION DID THIS STUDY ADDRESS?
This study characterized the pharmacokinetics (PKs) of sparsentan in healthy volunteers and patients with FSGS and evaluated the impact of clinically relevant disease characteristics and co‐medications as covariates.
WHAT DOES THIS STUDY ADD TO OUR KNOWLEDGE?
Sparsentan PKs were characterized by a two‐compartment model with first‐order absorption and lag time, dose‐dependent bioavailability, and first‐order elimination from the central compartment.
HOW MIGHT THIS CHANGE DRUG DISCOVERY, DEVELOPMENT, AND/OR THERAPEUTICS?
Patients taking strong CYP3A4 inhibitors concomitantly may require dose adjustments when adding sparsentan. However, other co‐medications and disease characteristics are not expected to require dose adjustments at this time.
INTRODUCTION
Focal segmental glomerulosclerosis (FSGS) is a rare, rapidly progressive kidney disease associated with kidney failure requiring dialysis or transplantation. It is the most common glomerular cause of end‐stage renal disease in children and adults in the United States. 1 , 2 , 3 FSGS is characterized by histological lesions resulting in segmental accumulation of glomerular extracellular matrix that leads to glomerular scarring and capillary obliteration due to podocyte injury‐triggered processes. FSGS can be classified as primary, genetic, secondary, or unexplained cause. Primary FSGS is the most common classification. 4 Signs and symptoms of primary FSGS include severe proteinuria, edema, and often fully developed nephrotic syndrome. 5 Current standard‐of‐care treatments for FSGS include off‐label treatment with angiotensin II receptor blockers (ARBs) and angiotensin‐converting enzyme (ACE) inhibitors. 6 Immunosuppressive treatments (corticosteroids and other immunosuppressants) are considered if patients do not respond to ARBs and ACE inhibitors. However, immunosuppressive drugs have serious adverse effects. 7
Sparsentan is a novel, first‐in‐class, single‐molecule dual endothelin angiotensin receptor antagonist (DEARA) being developed for the treatment of FSGS and IgAN. 8 Sparsentan is a highly selective antagonist for endothelin type A receptor (ETAR) and angiotensin II receptor type 1 (AT1R), which are both involved in mechanisms underlying the pathophysiology of rare proteinuric glomerular diseases. 9 , 10 , 11 , 12 , 13 , 14 Both endothelin‐1 and angiotensin II are vasoactive peptides that exert renal hemodynamic actions and promote cell growth, oxidative stress, and increased expression and activity of proinflammatory and profibrotic mediators. 15 , 16 , 17 , 18 , 19 , 20 , 21
Sparsentan has been investigated in healthy volunteers (phase I studies) as well as in patients with primary or genetic FSGS (phase II DUET 22 and phase III DUPLEX 23 studies). The phase II DUET study compared sparsentan (200, 400, or 800 mg/day) to irbesartan (300 mg/day) over 8 weeks. 22 Patients treated with sparsentan had a higher reduction in urinary protein to creatinine ratio (UP/C) and had higher rates of patients reaching the FSGS partial remission end point (FPRE). 22 , 24 The ongoing phase III DUPLEX study is investigating sparsentan treatment (400 mg/day for 2 weeks, titrating up to 800 mg/day) compared with irbesartan treatment (150 mg/day, titrating up to 300 mg/day) over 108 weeks. 23 The primary end point of the DUPLEX study is estimated glomerular filtration rate slope from week 6 to week 108, and the prespecified interim surrogate end point is the proportion of patients achieving FPRE (UP/C ≤1.5 g/g and >40% reduction in UP/C) at week 36. 23 Interim results showed that sparsentan treatment led to significantly greater FPRE response compared to irbesartan. 23
A population pharmacokinetic (PK) analysis was conducted and reported to characterize the PKs of sparsentan in healthy volunteers and patients with primary or genetic FSGS and to evaluate the impact of FSGS disease characteristics and concomitant medications on sparsentan PKs.
METHODS
Subjects, PK serum sampling, and bioanalysis
PK data from nine studies, including six phase I studies in healthy volunteers, one phase I study comparing subjects with hepatic impairment and healthy volunteers, and two studies in patients with primary and genetic FSGS (phase II DUET and phase III DUPLEX; Table S1) were collected after sparsentan treatment. All studies were approved by independent ethics committees and conducted in accordance with the Declaration of Helsinki. All patients were provided written informed consent.
Sparsentan plasma concentrations were determined using the validated liquid chromatography–tandem mass spectrometry method. Sparsentan and sparsentan‐d5 were extracted as internal standards in K2EDTA plasma samples by protein precipitation with acetonitrile. The sample supernatant was evaporated to dryness and reconstituted with 5 mM ammonium acetate in H2O:MeCN (70:30 v/v). The chromatographic separation was achieved with Atlantis d18 (3 μm, 2.1 × 50 mm; Waters) using mobile phase solvent A (5 mM ammonium acetate in water), solvent B (acetonitrile), and solvent C (H2O:MeCN [50:50 v/v]) at a flow rate of 0.5 mL/min (pump C = 0.3 mL/min). The calibration range was 2.00 to 4000 ng/mL for sparsentan with a lower limit of quantitation (LLOQ) of 2 ng/mL. PK samples of subjects with at least one sparsentan dose administration with at least one associated postdose plasma sparsentan concentration above the LLOQ were included, and all PK samples below the LLOQ were excluded from the analysis.
Development of the population PK model
The population PK analysis was performed using nonlinear mixed‐effects modeling in NONMEM (version 7.5; ICON Development Solutions) with the first‐order conditional estimation with η–ϵ interaction (FOCE‐1) method. Perl‐Speak‐NONMEM (version 5.0; Uppsala University, Uppsala, Sweden) was used to facilitate the covariate search and the evaluation and validation of the PK models, and the results were further analyzed by R (version 4.0 or higher).
To establish the base model, one‐ and two‐compartment models were evaluated based on the shape of the observed concentration‐time profile. Zero‐order absorption processes and lag times followed by first‐order absorption processes were also explored. Based on the study result of study 021HVOL16008 (effect of sparsentan on midazolam; data on file) and study RTRX‐RE021‐103 (multiple ascending dose study of sparsentan; data on file), time‐dependent clearance (CL) due to CYP3A auto‐induction was examined as follows: CL(t) = CL0 + ΔCL × (1 − exp[−k induction × day]), where CL(t) is time‐varying CL, CL0 is the initial CL for the first dose, ΔCL is the additional CL at steady‐state, and k induction is the rate of induction onset. Interindividual random effects on the parameters were introduced and retained if they did not cause model instability and if the estimates were not close to zero. They were modeled assuming a log‐normal distribution as given by the following:
where θ i denotes the parameter value for the i th subject, θ Typical denotes the typical parameter value, and η i denotes the interindividual random effect for the i th subject, assumed to be normally distributed with mean 0 and variance ω. 2
A total of 20 covariates were considered and tested based on their clinical relevance to greater than or equal to one population PK parameters, including standard demographics, laboratory‐based liver function markers, creatinine clearance (CrCL; Cockcroft‐Gault formula), food status, formulation (capsule and tablet), population (FSGS and healthy volunteers), and concomitant medications (acid‐reducing agent, CYP3A inducer and inhibitor, and P‐glycoprotein [P‐gp] inhibitor; Table S2). The median value of the study population was used for missing continuous covariate data points, whereas missing categorical covariates were flagged without inputting a data point. Covariates missing data in greater than 20% of the population were not included in the population analysis. Covariates were tested in a stepwise process, initially examined using pairwise plots of random effects versus covariates using results of the base model. Then, covariates were added one at a time. All covariates p < 0.01 were entered into a full model. Stepwise backward elimination was applied to the full model, deleting covariates in inverse order of significance until all remaining covariates were significant at p < 0.001. Significance levels were evaluated using a log‐likelihood criterion. Covariates were added in the log‐domain. Continuous covariates were incorporated into the population model using a scaled structure based on the median (population) or standard value of the covariate (Table 1). Categorical covariates were incorporated into the population model using the most frequent level of the covariate as the reference (Table 1).
TABLE 1.
Summary of baseline continuous and categorical covariates (overall subjects, n = 446).
| Covariate | Healthy subjects, n = 236 | FSGS subjects, n = 194 | Subjects with hepatic impairment, n = 16 | Overall, n = 446 |
|---|---|---|---|---|
| Age, years | ||||
| Mean (SD) | 38.5 (9.7) | 41.0 (16.9) | 56.8 (5.7) | 40.3 (13.6) |
| Median (CV%) | 38.0 (25.1) | 43.0 (41.2) | 58.0 (10.0) | 40.0 (33.9) |
| Range | 18.0–65.0 | 8.0–74.0 | 49.0–65.0 | 8.0–74.0 |
| Body weight, kg | ||||
| Mean (SD) | 79.8 (12.4) | 80.6 (21.8) | 82.2 (14.1) | 80.2 (17.2) |
| Median (CV%) | 78.0 (15.6) | 79.3 (27.1) | 84.7 (17.2) | 78.6 (21.4) |
| Range | 54.6–122.0 | 21.1–154.0 | 53.0–103 | 21.1–154.0 |
| BSA (m2) | ||||
| Mean (SD) | 1.92 (0.195) | 1.90 (0.289) | 1.95 (0.178) | 1.91 (0.240) |
| Median (CV%) | 1.92 (10.1) | 1.89 (15.2) | 1.98 (9.1) | 1.91 (12.5) |
| Range | 1.42–2.51 | 0.83–2.61 | 1.59–2.17 | 0.83–2.61 |
| Missing, n (%) | – | 1 (0.5) | – | 1 (0.2) |
| BMI, kg/m2 | ||||
| Mean (SD) | 26.9 (2.79) | 28.1 (6.18) | 27.9 (4.11) | 27.5 (4.65) |
| Median (CV%) | 27.0 (10.4) | 27.7 (22.0) | 27.8 (14.8) | 27.4 (16.9) |
| Range | 19.1–34.0 | 15.1–47.0 | 18.8–34.4 | 15.1–47.0 |
| Missing, n (%) | – | 1 (0.5) | – | 1 (0.2) |
| Lean body weight, kg | ||||
| Mean (SD) | 58.4 (9.3) | 55.3 (12.0) | 60.5 (7.0) | 57.1 (10.6) |
| Median (CV%) | 59.0 (15.9) | 54.0 (21.7) | 62.2 (11.5) | 57.7 (18.5) |
| Range | 35.8–83.1 | 18.5–84.8 | 45.6–68.8 | 18.5–84.8 |
| Missing, n (%) | – | 1 (0.5) | – | 1 (0.2) |
| Albumin, g/dL | ||||
| Mean (SD) | 4.40 (0.359) | 3.50 (0.727) | 4.12 (0.574) | 4.00 (0.711) |
| Median (CV%) | 4.40 (8.1) | 3.70 (20.8) | 4.25 (13.9) | 4.10 (17.8) |
| Range | 3.40–5.50 | 1.40–4.80 | 3.00–5.00 | 1.40–5.50 |
| Total protein, g/dL | ||||
| Mean (SD) | 7.23 (0.50) | 5.82 (1.03) | 7.38 (0.73) | 6.62 (1.05) |
| Median (CV%) | 7.25 (6.9) | 5.90 (17.7) | 7.60 (9.8) | 6.90 (15.9) |
| Range | 5.60–8.80 | 3.20–7.90 | 5.90–8.70 | 3.20–8.80 |
| CrCL, mL/min | ||||
| Mean (SD) | 127 (26.9) | 90.2 (43.7) | 128 (29.7) | 111 (39.7) |
| Median (CV%) | 123 (21.3) | 78.4 (48.5) | 126 (23.3) | 112 (35.8) |
| Range | 73.8–253.0 | 26.0–361.0 | 87.4–189.0 | 26.0–361.0 |
| Creatinine, μmol/L | ||||
| Mean (SD) | 79.1 (16.6) | 115 (46.8) | 67.8 (14.2) | 94.5 (38.0) |
| Median (CV%) | 79.6 (21.0) | 111 (40.5) | 66.3 (21.0) | 86.6 (40.2) |
| Range | 41.5–122.0 | 35.4–240.0 | 44.2–97.2 | 35.4–240.0 |
| AST, U/L | ||||
| Mean (SD) | 21.5 (6.0) | 23.0 (11.5) | 63.8 (50.0) | 23.7 (14.9) |
| Median (CV%) | 21.0 (27.9) | 20.0 (49.9) | 37.0 (78.5) | 21.0 (62.8) |
| Range | 11.0–43.0 | 8.0–95.0 | 17.0–167 | 8.0–167.0 |
| ALT, U/L | ||||
| Mean (SD) | 22.0 (10.3) | 21.3 (12.0) | 60.6 (52.4) | 23.1 (16.3) |
| Median (CV%) | 19.0 (46.9) | 18.0 (56.3) | 38.5 (86.6) | 19.0 (70.4) |
| Range | 5.0–59.0 | 5.0–82.0 | 14.0–163.0 | 5.0–163.0 |
| Missing, n (%) | 1 (0.4) | – | – | 1 (0.2) |
| Total bilirubin, μmol/L | ||||
| Mean (SD) | 9.64 (5.63) | 6.41 (3.88) | 21.9 (14.60) | 8.67 (6.29) |
| Median (CV%) | 8.60 (58.4) | 5.10 (60.5) | 18.8 (66.7) | 6.80 (72.5) |
| Range | 1.7–46.2 | 1.7–27.0 | 5.10–51.3 | 1.7–51.3 |
| Missing, n (%) | 3 (1.3) | – | – | 3 (0.7) |
| ALKP, U/L | ||||
| Mean (SD) | 67.2 (19.8) | 79.6 (36.2) | 95.1 (35.0) | 73.6 (29.5) |
| Median (CV%) | 65.0 (29.5) | 73.0 (45.4) | 91.5 (36.8) | 68.0 (40.1) |
| Range | 30.0–174.0 | 25.0–269.0 | 47.0–165.0 | 25.0–269.0 |
| Sex, n (%) | ||||
| Female | 62 (26.3) | 86 (44.3) | 0 (0) | 148 (33.2) |
| Male | 174 (73.7) | 108 (55.7) | 16 (100) | 298 (66.8) |
| Race, n (%) | ||||
| White | 141 (59.7) | 146 (75.3) | 14 (87.5) | 301 (67.5) |
| Black or African American | 89 (37.7) | 15 (7.7) | 1 (6.2) | 105 (23.5) |
| Asian | 4 (1.7) | 22 (11.3) | 1 (6.2) | 27 (6.1) |
| Multiple | 2 (0.8) | 1 (0.5) | 0 (0) | 3 (0.7) |
| Other | 0 (0) | 10 (5.2) | 0 (0) | 10 (2.2) |
| Renal function, n (%) | ||||
| Normal | 231 (97.9) | 85 (43.8) | 15 (93.8) | 331 (74.2) |
| Mild | 5 (2.1) | 54 (27.8) | 1 (6.2) | 60 (13.5) |
| Moderate | 0 (0) | 53 (27.3) | 0 (0) | 53 (11.9) |
| Severe | 0 (0) | 2 (1.0) | 0 (0) | 2 (0.4) |
| Hepatic function, n (%) | ||||
| Normal | 12 (5.1) | 0 (0) | 0 (0) | 12 (2.7) |
| Mild | 0 (0) | 0 (0) | 8 (50.0) | 8 (1.8) |
| Moderate | 0 (0) | 0 (0) | 8 (50.0) | 8 (1.8) |
| Severe | 0 (0) | 0 (0) | 0 (0) | 0 (0) |
| Missing | 224 (94.9) | 194 (100) | – | 418 (93.7) |
| Food, n (%) a | ||||
| Fasting | 236 (100) | 0 (0) | 16 (100) | 252 (56.5) |
| Fed | 51 (21.6) | 0 (0) | 0 (0) | 51 (11.4) |
| Unknown | 0 (0) | 194 (100) | 0 (0) | 194 (43.5) |
| Formulation, n (%) a | ||||
| Capsule | 137 (58.1) | 71 (36.6) | 0 (0) | 208 (46.6) |
| Tablet | 95 (40.3) | 123 (63.4) | 16 (100) | 234 (52.5) |
| Crushed tablet | 36 (15.3) | 0 (0) | 0 (0) | 36 (8.1) |
| P‐gp inhibitor, n (%) | ||||
| No co‐administration | 220 (93.2) | 174 (89.7) | 16 (100) | 410 (91.9) |
| Co‐administration | 16 (6.8) | 20 (10.3) | 0 (0) | 36 (8.1) |
| CYP3A4 inhibitor, n (%) a | ||||
| None | 236 (100) | 118 (60.8) | 16 (100) | 370 (83.0) |
| Unknown | 0 (0) | 0 (0) | 0 (0) | 0 (0) |
| Weak | 0 (0) | 57 (29.4) | 0 (0) | 57 (12.8) |
| Moderate | 30 (12.7) | 17 (8.8) | 0 (0) | 47 (10.5) |
| Strong | 30 (12.7) | 2 (1.0) | 0 (0) | 32 (7.2) |
| CYP3A4 inducer, n (%) | ||||
| None | 236 (100) | 140 (72.2) | 16 (100) | 392 (87.9) |
| Unknown | 0 (0) | 54 (27.8) | 0 (0) | 54 (12.1) |
| Weak | 0 (0) | 0 (0) | 0 (0) | 0 (0) |
| Moderate | 0 (0) | 0 (0) | 0 (0) | 0 (0) |
| Strong | 0 (0) | 0 (0) | 0 (0) | 0 (0) |
| Acid‐reducing agent, n (%) | ||||
| No co‐administration | 236 (100) | 135 (69.6) | 16 (100) | 387 (86.8) |
| Co‐administration | 0 (0) | 59 (30.4) | 0 (0) | 59 (13.2) |
Abbreviations: ALKP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; BSA, body surface area; CrCL, creatinine clearance; CV, coefficient variation; CYP, cytochrome P450; FSGS, focal segmental glomerulosclerosis; P‑gp, P‑glycoprotein.
In some cases, the n per category exceeds the total n. This is because some studies had cross‑over periods, and subjects were counted in multiple categories.
The base and final models were evaluated based on reduction in minimum objective function value (OFV), visual inspection of diagnostic plots, plausibility of parameter estimates and their relative standard error values, change in OFV relative to the change in the number of parameters, and changes in both interindividual and residual variability. A prediction‐corrected visual predictive check (VPC) was performed to compare distributions of simulated sparsentan concentrations from the final model and distributions of the observed sparsentan concentrations. In addition, a nonparametric bootstrap resampling analysis was performed to evaluate the stability of the final model and to estimate standard errors for the model parameters. 25
Model application
The final model was used to quantify the univariate and multivariate effects of covariates on the expected steady‐state exposure of sparsentan in subjects with FSGS receiving 800 mg/day, including area under the concentration‐time curve (AUC), maximum concentration (C max), and minimum or trough concentration (C min).
To quantify the univariate effects of covariates on the exposures, the final model was used to simulate the sparsentan exposure for a typical subject (a white man receiving an 800‐mg tablet with median alkaline phosphatase [ALKP] and CrCL values and not receiving a moderate or strong CYP3A4 inhibitor) with specified settings for significant model covariates. Continuous covariates were varied one‐at‐a‐time to 5th and 95th percentile values, and categorical covariates were varied to different category values. To quantify the multivariate effects of covariates on exposures, individual exposures using each subject's estimated PK parameters were stratified post hoc by covariate category.
RESULTS
Participants
After exclusions, the dataset for the final model included 10,957 samples from 446 subjects (including 1683 samples from 194 patients with FSGS) in nine studies (Table S1). Baseline demographic and clinical characteristics of the subjects included in the population PK analysis are shown in Table 1.
Base model
A two‐compartment model with first‐order absorption and an absorption lag time (T lag), with proportional plus additive residual error was used for the base model. To account for the dose‐dependent bioavailability, a nonlinear relationship on relative bioavailability (F rel) was used for doses above 200 mg to account for dose‐dependent bioavailability. To account for lower plasma concentrations of sparsentan at steady‐state than following a single dose, an induction term on apparent clearance (CL/F) modeled as a rapid increase to steady‐state occurring after the first dose during multiple‐dose regimens was used. The onset of induction was set with a half‐life of 0.001 days arbitrarily because the time course was not apparent in C min. This assumption was tested, and the estimated half‐life was 0.3 days, justifying the arbitrary setting.
Final model
Starting from the base model, a univariate forward‐stepwise backward covariate analysis was performed. A total of 20 covariate‐parameter relationships were significant in the univariate forward search at a significance level of p < 0.01 and were included in the full model. In the backward elimination step at p < 0.001, six covariate‐parameter relationships were retained: ALKP, CrCL, and sex on CL; race on central volume of distribution (V c); formulation on T lag; and formulation on absorption rate constant (K a; Table S3). The effect of each covariate on change in OFVs is presented in Table S4.
In the final population PK model, CL/F was 3.88 L/h after a single 400‐mg dose, increasing to 5.12 L/h at steady‐state. The apparent central volume of distribution (V c/F) and apparent peripheral volume of distribution (V p/F) estimates were 49.3 and 12.1 L, respectively. The K a and T lag were 0.740 h and 0.32 h, respectively. The terminal half‐life was 9.6 h at steady‐state (Table 2). After a single‐800 mg dose, the CL/F was 5.47 L/h, increasing to 7.21 L/h at steady‐state. The V c/F and V p/F estimates were 69.5 and 17.0 L, respectively. Given the final population PK model structure, the terminal half‐life after 800‐mg dose remained the same as that of a 400‐mg dose at steady‐state. Dose nonlinearity parameter on F rel is −0.495, suggesting less than dose‐proportional exposures. As a result, F rel was estimated at 1.41, 1.00, and 0.71 for doses of 200, 400, and 800 mg, respectively.
TABLE 2.
Final model parameter estimates.
| Parameter | Estimate | RSE (%) | IIV (%) | Shrinkage (%) |
|---|---|---|---|---|
| CL/F (L/h) | 3.88 | 4.6 | 39.5 | 4.7 |
| V c/F (L) | 49.3 | 4.3 | 48.4 | 17.2 |
| Q/F (L/h) | 2.03 | 12.0 | – | – |
| V p/F (L) | 12.1 | 10.5 | – | – |
| K a (1/h) | 0.740 | 6.9 | 68.9 | 22.2 |
| T lag (h) | 0.32 | 4.0 | NA | NA |
| T 1/2 (h) (derived) | 9.6 | NA | NA | NA |
| Induction change in CL (L/h) | 1.23 | 13.6 | NA | NA |
| Induction t 1/2 (day) | 0.001 | Fixed a | NA | NA |
| Dose on F rel | −0.495 | 5.1 | NA | NA |
| Effect on CL | ||||
| Moderate CYP3A4 | −0.273 | 18.8 | NA | NA |
| Strong CYP3A4 | −1.069 | 10.0 | NA | NA |
| ALKP | −0.208 | 27.5 | NA | NA |
| CrCL | 0.222 | 26.5 | NA | NA |
| Maleb | 0.139 | 32.8 | NA | NA |
| Effect on V c b | ||||
| Black or African American | 0.309 | 18.4 | NA | NA |
| Asian | 0.265 | 48.4 | NA | NA |
| Effect on K a | ||||
| Tablet | −0.306 | 34.8 | NA | NA |
| Crushed tablet | 0.080 | 159.1 | NA | NA |
| Effect on T lag | ||||
| Tablet | −0.269 | 29.1 | NA | NA |
| Crushed tablet | −1.175 | 28.7 | NA | NA |
| Variance CL | 0.156 | 8.7 | NA | NA |
| Variance V c | 0.234 | 11.3 | NA | NA |
| Variance K a | 0.474 | 10.2 | NA | NA |
| SD of additive error (ng/mL) | 2 | Fixed a | NA | NA |
| SD of proportional error | 0.365 | 1.9 | NA | NA |
| 800 mg doseb | ||||
| CL/F (L/h) (derived) | 5.47 | – | NA | NA |
| Steady‐state CL (L/h) (derived) | 7.21 | – | NA | NA |
| V c /F (L) (derived) | 69.5 | – | NA | NA |
| V p /F (L) (derived) | 17.0 | – | NA | NA |
| T 1/2 (h) (derived) | 9.6 | NA | NA | NA |
Note: F rel = (dose/400)−0.495, if dose ≥200 mg; F = (200/400)−0.495, if dose <200 mg.
Abbreviations: ALKP, alkaline phosphatase; CL, clearance; CL/F, apparent clearance; CrCL, creatinine clearance; CYP, cytochrome P450; F, bioavailability; F rel, relative bioavailability; IIV, interindividual variance; K a, absorption rate constant; Q/F, apparent distribution clearance; RSE, relative standard error; t 1/2, half‐life; T lag, absorption lag time; V c, central volume of distribution; V c/F, apparent central volume; V p, peripheral volume of distribution; V p/F, apparent peripheral volume.
Induction t 1/2 and SD of additive error are fixed and not estimated.
The reference subject is a white female receiving a 400‐mg capsule, no CYP3A4 inhibitor, with CrCL of 112 mL/min, and with ALKP of 68 U/L.
Increased ALKP and decreased CrCL both correlated to decreased CL. Men had higher CL than women, and Black subjects had a higher V c than White subjects. Relative to the capsule, the tablet and crushed tablet had a shorter T lag, and the tablet was absorbed more slowly.
Goodness‐of‐fit plots indicate that no bias is observed (Figure S1). VPCs after a single dose and at steady‐state in healthy volunteers, subjects with hepatic impairment, and patients with FSGS are shown in Figure 1.
FIGURE 1.
Visual predictive checks after a single dose in the following (a) healthy volunteers and volunteers with hepatic impairment; (b) in healthy volunteers at steady state; (c) patients with FSGS following first dose; and (d) patients with FSGS at steady‐state. Dots are prediction‐corrected concentrations. Gray shaded area shows the 5th to 95th percentile range of the model prediction. Blue lines show the 5th, median, and 95th percentile range of the model prediction. Gold dashed lines show the 5th, median, and 95th percentile range of the concentrations. FSGS, focal segmental glomerulosclerosis.
Impact of covariates on sparsentan exposure
The univariate effect of statistically significant covariates on sparsentan PK parameters compared to the typical subject (white male subject, 800‐mg tablet, no moderate or strong CYP3A4 inhibitors) can be found in Table 3 and Figure 2. Using these parameters, formulation had a minimal effect on the steady‐state PK profile, with AUC unaffected and C max increased by less than 10% for both the capsule and crushed tablet formulations compared to tablet formulation (Figure 3). Strong CYP3A4 inhibitor comedications increased AUC and C max by 191.3% and 99.0%, respectively. Moderate CYP3A4 inhibitor comedications increased AUC and C max by 31.4% and 16.0%, respectively. Other covariates affected AUC and C max by less than 20%.
TABLE 3.
Influence of covariates on FSGS PK parameters.
| PK parameters and baseline covariates | Percentile | Baseline covariate value | Estimate | Percent change from typical (%) |
|---|---|---|---|---|
| Typical CL/F (L/h) | 7.5 | |||
| Alkaline phosphatase (U/L) | 5th | 40 | 8.5 | 13 |
| 95th | 140 | 6.6 | −13 | |
| Creatinine clearance, mL/min | 5th | 41 | 6.5 | −13 |
| 95th | 164 | 8.9 | 18 | |
| Sex | – | Female | 6.6 | −13 |
| CYP3A4 inhibitor | – | Moderate | 5.7 | −24 |
| – | Strong | 2.6 | −66 | |
| Typical V c /F (L) | 69 | |||
| Race | – | Black | 95 | 36 |
| – | Asian | 91 | 30 | |
| Lag time (h) | 0.24 | |||
| Formulation | – | Capsule | 0.32 | 31 |
| – | Crushed tablet | 0.099 | −60 | |
| Absorption rate constant (1/h) | 0.54 | |||
| Formulation | – | Capsule | 0.74 | 36 |
| – | Crushed tablet | 0.8 | 47 |
Note: The typical FSGS subject is a white male receiving an 800‐mg tablet with ALKP of 73 U/L, CrCL of 78 mL/min, and not receiving a moderate or strong CYP3A4 inhibitor.
Abbreviations: ALKP, alkaline phosphatase; CL/F, apparent clearance; CYP, cytochrome P450; FSGS, focal segmental glomerulosclerosis; PK, pharmacokinetic; V c /F, apparent central volume of distribution.
FIGURE 2.
Tornado plots showing the effects of covariates on C min (left), steady‐state AUC (center), and C max (right). Base, as represented by the black vertical line, refers to the predicted steady‐state AUC of sparsentan in a typical male subject with FSGS receiving an 800 mg tablet with median ALKP of 73 U/L, with median CrCL of 78 mL/min, and not receiving a moderate or strong CYP3A4 inhibitor. Blue shaded bar shows the 5th to 95th percentile exposure range in the FSGS population. Green shaded bars represent the influence of a single covariate on the steady‐state exposure after once‐daily sparsentan 800 mg. Green bars are ranked in decreasing order of largest deviation from the base. Upper and lower values for each covariate capture 90% of the range in the population. ALKP, alkaline phosphatase; AUC, area under the plasma concentration‐time curve; C max, maximum concentration; C min, minimum or trough concentration; CrCL, creatinine clearance; CYP, cytochrome P450; FSGS, focal segmental glomerulosclerosis.
FIGURE 3.
Modeled steady‐state FSGS PK profiles stratified by formulation. Simulations use parameters for a white male receiving 800 mg with ALKP of 73 U/L, with CrCL of 78 mL/min, and not receiving a moderate or strong CYP3A4 inhibitor. ALKP, alkaline phosphatase; CrCL, creatinine clearance; CYP, cytochrome P450; FSGS, focal segmental glomerulosclerosis; PK, pharmacokinetic.
Model application: Exposure among various FSGS subpopulations
An increase in sparsentan steady‐state exposure was shown with increasing age brackets (8–17, 18–64, and ≥65 years; Figure S2), although age was not a significant covariate in the final population PK model. Age is an input for CrCL, and the inclusion of CrCL in the final model likely contributes to this result. Exposure also largely overlaps among the age groups, and thus, no dose adjustment based on age is considered necessary. Sparsentan exposures at steady‐state were comparable across race (Figure S3); thus, dose adjustment based on race is not justified.
DISCUSSION
This population PK analysis of sparsentan included 10,957 plasma samples collected in nine clinical studies involving 236 healthy volunteers, 16 subjects with hepatic impairment compared to healthy subjects, and 194 patients with primary and genetic FSGS. The PKs of sparsentan were adequately described by a two‐compartment model with first‐order absorption and a lag time, dose‐dependent bioavailability, and first‐order elimination from the central compartment. A modest increase in CL at steady‐state was incorporated. At the intended daily dose of 800 mg, CL/F of sparsentan was estimated at 5.47 L/h after a single dose, increasing to 7.21 L/h at steady‐state. This may account for the decrease in exposures observed in the DUET study between patients' single‐dose sample and steady‐state samples. The model suggests that dose adjustments may only be needed sparingly for sparsentan.
Sparsentan is primarily metabolized by CYP3A and has been shown to be a competitive inhibitor and inducer of CYP3A in vitro (data on file). Therefore, the exposure seen at steady‐state is possibly due to CYP3A auto‐induction. In the exploratory data analysis, auto‐induction appeared to be stronger at the higher doses. However, the dose‐dependent auto‐induction was tested in the base PK models and was found not to be significant. The effect was modeled in the final PK model as occurring instantly after the first dose, whereas the CYP3A induction could take days to weeks. 26 In the base PK models, a time‐varying model was tested and estimated the CL change to occur with an equilibration half‐life of 0.2 days. However, the result should be interpreted with caution as the current sparsentan PK data might not allow precise estimation of the time frame.
In the covariate analysis, formulation had a significant effect on the K a and lag time. The tablet was absorbed more slowly than the capsule, whereas the crushed tablet was absorbed more quickly than the whole tablet or the capsule. The change in C max was less than 10%, with no overall change in the AUC because there was no effect on F rel. Thus, the clinical effect of the formulation on the PK profile was negligible.
The strongest covariates were the impact of moderate and strong CYP3A4 inhibitors, associated with increases of 31% and 191%, respectively, in the steady‐state AUC. The effect of CYP3A4 inhibitors was overall consistent with observed sparsentan exposures in healthy volunteers when co‐administrating with cyclosporine (a moderate CYP3A inhibitor; 70% increase in sparsentan AUC to infinity [AUCinf]) and itraconazole (a strong CYP3A inhibitor; 174% increase in sparsentan AUCinf; study 021HVOL16006, data on file).
Decreased CrCL was associated with increased sparsentan exposures. Given that urinary excretion is a minor elimination route for sparsentan (2.18% of the radioactive dose in the mass balance study, study 021HVOL16005, data on file), this relationship is recognized but not fully understood. The calculation of CrCL based on the Cockcroft‐Gault equation requires the variable of sex, which is also a covariate on CL/F. Thus, the apparent association might be confounded by sex. In addition, it has been shown that nonrenal elimination pathways may be altered by renal impairment through indirect or secondary inhibitory effects of elevated uremic toxins on metabolizing CYP enzymes. 27 Nevertheless, the magnitude of the impact of CrCL at the 5th and 95th percentile on sparsentan exposure is relatively small (<20% on AUC and <10% on C max), and exposure variability is not expected to have a clinically meaningful impact on safety for patients with mild and moderate renal impairment. However, no conclusions can be drawn regarding the effects of severe renal impairment due to limited number of patients (n = 2).
Hepatic impairment was not directly tested as a covariate. Instead, several hepatic markers were tested, including aspartate aminotransferase (SGOT/AST), alanine aminotransferase (SGPT/ALT), ALKP, and total bilirubin. Overall, these covariates showed no relationship to CL/F, except for ALKP, where elevated ALKP was associated with a decreased CL and elevated exposures. However, the effect of ALKP at the 5th and 95th percentile on exposure is small (<15% on AUC and <10% on C max) and does not appear to translate into clinically relevant differences in safety, which is consistent with the result from the dedicated phase I study with mild and moderate hepatic impairment not having a large effect on sparsentan exposure (study 021IHFX16009, data on file).
Female subjects had higher exposure, and race (White, Black, and Asian) significantly impacted the V c /F and PK variability. However, the impact of these covariates on sparsentan steady‐state exposure was small (<15% on AUC and C max), and thus, no dose adjustment based on sex or race is considered necessary.
Although the impact of each significant covariate was relatively small, except for the effect of the strong CYP3A inhibitor, an additional simulation was performed to evaluate the combined effect of the covariates. For a White, female, patient with FSGS and with CrCL at 5th percentile (41 mL/min), ALKP at the 95th percentile (141 U/L), and with concomitant use of a moderate CYP3A4 inhibitor, the steady‐state AUC after daily 800 mg sparsentan would be 193 μg/mL*h, representing an 82% increase in exposure. Although the combination of covariates may predispose a patient toward an exposure increase of greater than 50%, the patients represented by the simulation account for roughly 0.2% of the study population.
In the majority of phase I studies in healthy volunteers, sparsentan was administered with overnight fasting. In the phase II DUET and phase III DUPLEX studies, patients with FSGS were instructed to take sparsentan before the first meal of the day. In the population PK analysis, there was no significant impact on absorption parameters between dosing before the first meal of the day in patients with FSGS compared to dosing with overnight fasting in healthy subjects.
Other intrinsic covariates, including population (patients with FSGS vs. healthy volunteers), age, weight, SGOT/AST, SGPT/ALT, total bilirubin, albumin, and total protein had no significant effect on the CL/F or apparent volume of distribution of sparsentan. No significant effect of co‐medications with acid reducer (including proton pump inhibitors), P‐gp inhibitor, or mild CYP3A4 inhibitor was detected, suggesting sparsentan can be co‐administered safely with these agents without dose adjustment.
The population PK model accurately characterized sparsentan exposure in healthy volunteers and patients with primary and genetic FSGS. The population PK analysis suggests that dose adjustments for sparsentan may be warranted for patients taking moderate or strong CYP3A inhibitors concomitantly, but no adjustments are needed across the range of other concomitant medications, patient characteristics, and disease characteristics analyzed.
AUTHOR CONTRIBUTIONS
R.W., H.J.K., L.Z., and S.C.C. wrote the manuscript. R.W. and S.C.C. designed the research. R.W., H.J.K., and L.Z. performed the research. R.W. and S.C.C. analyzed the data.
FUNDING INFORMATION
This study was funded by Travere Therapeutics, Inc (San Diego, CA). Travere Therapeutics, Inc., provided financial and material support for the research. Investigators were paid consultants (R.W., H.J.K., and L.Z.) or an employee and stockholder (S.C.C.) of Travere Therapeutics, Inc.
CONFLICT OF INTEREST STATEMENT
R.W. is currently with QuanTx Consulting; was formerly with Certara; and is a paid consultant to Travere Therapeutics, Inc., in connection with this work. H.J.K. and L.Z. are employees of Certara and are paid consultants to Travere Therapeutics, Inc., in connection with this work. S.C.C. is a former employee of and may have equity ownership in Travere Therapeutics, Inc.
PRIOR PRESENTATION/PUBLICATION
This study was previously presented at the American Society for Clinical Pharmacology & Therapeutics 2022 Annual Meeting, March 16–18, 2022, as Poster P‐036.
Supporting information
Figure S1
Figure S2
Figure S3
Table S1
Table S2
Table S3
Table S4
Table S5
ACKNOWLEDGMENTS
The authors thank Mari Willeman, PhD, Courtney Breuel, ELS, and Stephen Bublitz, ELS, of MedVal Scientific Information Services, LLC (Princeton, NJ), for medical writing and editorial assistance, which were funded by Travere Therapeutics, Inc. This manuscript was prepared according to the International Society for Medical Publication Professionals' “Good Publication Practice for Communicating Company‐Sponsored Medical Research: GPP3.”
Wada R, Kleijn HJ, Zhang L, Chen S‐C. Population pharmacokinetic analysis of sparsentan in healthy volunteers and patients with focal segmental glomerulosclerosis. CPT Pharmacometrics Syst Pharmacol. 2023;12:1080‐1092. doi: 10.1002/psp4.12996
DATA AVAILABILITY STATEMENT
The data used for the analyses in this manuscript are available on request from the corresponding author.
REFERENCES
- 1. Fogo AB. Causes and pathogenesis of focal segmental glomerulosclerosis. Nat Rev Nephrol. 2015;11(2):76‐87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Korbet SM. Treatment of primary FSGS in adults. J Am Soc Nephrol. 2012;23(11):1769‐1776. [DOI] [PubMed] [Google Scholar]
- 3. United States Renal Data System . ESRD Among Children, Adolescents, and Young Adults. USRDS 2021 Annual data report: epidemiology of kidney Disease in the United States, volume 2, Chapter 8. 2021. https://www.usrds.org/media/1627/v2_c08_pediatric_16.pdf
- 4. D'Agati VD, Kaskel FJ, Falk RJ. Focal segmental glomerulosclerosis. N Engl J Med. 2011;365(25):2398‐2411. [DOI] [PubMed] [Google Scholar]
- 5. Rosenberg AZ, Kopp JB. Focal segmental glomerulosclerosis. Clin J Am Soc Nephrol. 2017;12(3):502‐517. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Campbell KN, Pennese N, Zaffalon A, et al. Efficacy and safety of ACE inhibitor and angiotensin receptor blocker therapies in primary focal segmental glomerulosclerosis treatment: a systematic review and meta‐analysis. Kidney Med. 2022;4(5):100457. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Kidney Disease: Improving global outcomes (KDIGO) glomerular diseases work group. KDIGO 2021 clinical practice guideline for the management of glomerular diseases. Kidney Int. 2021;100(4s):S1‐S276. [DOI] [PubMed] [Google Scholar]
- 8. Komers R, Plotkin H. Dual inhibition of renin‐angiotensin‐aldosterone system and endothelin‐1 in treatment of chronic kidney disease. Am J Physiol Regul Integr Comp Physiol. 2016;310(10):R877‐R884. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Gerstung M, Roth T, Dienes HP, Licht C, Fries JW. Endothelin‐1 induces NF‐kB via two independent pathways in human renal tubular epithelial cells. Am J Nephrol. 2007;27(3):294‐300. [DOI] [PubMed] [Google Scholar]
- 10. Lenoir O, Milon M, Virsolvy A, et al. Direct action of endothelin‐1 on podocytes promotes diabetic glomerulosclerosis. J Am Soc Nephrol. 2014;25(5):1050‐1062. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Watson AM, Li J, Schumacher C, et al. The endothelin receptor antagonist avosentan ameliorates nephropathy and atherosclerosis in diabetic apolipoprotein E knockout mice. Diabetologia. 2010;53(1):192‐203. [DOI] [PubMed] [Google Scholar]
- 12. Junaid A, Rosenberg ME, Hostetter TH. Interaction of angiotensin II and TGF‐b1 in the rat remnant kidney. J Am Soc Nephrol. 1997;8(11):1732‐1738. [DOI] [PubMed] [Google Scholar]
- 13. Kashiwagi M, Shinozaki M, Hirakata H, et al. Locally activated renin‐angiotensin system associated with TGF‐b1 as a major factor for renal injury induced by chronic inhibition of nitric oxide synthase in rats. J Am Soc Nephrol. 2000;11(4):616‐624. [DOI] [PubMed] [Google Scholar]
- 14. Yang F, Chung AC, Huang XR, Lan HY. Angiotensin II induces connective tissue growth factor and collagen I expression via transforming growth factor‐b‐dependent and ‐independent Smad pathways: the role of Smad3. Hypertension. 2009;54(4):877‐884. [DOI] [PubMed] [Google Scholar]
- 15. Saleh MA, Boesen EI, Pollock JS, Savin VJ, Pollock DM. ETA receptor specific stimulation of glomerular inflammation and injury in streptozotocin‐induced diabetic rats. Diabetologia. 2011;54(4):979‐988. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Saleh MA, Boesen EI, Pollock JS, Savin VJ, Pollock DM. Endothelin‐1 increases glomerular permeability and inflammation independent of blood pressure in the rat. Hypertension. 2010;56(5):942‐949. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Saleh MA, Pollock JS, Pollock DM. Distinct actions of endothelin A‐selective versus combined endothelin a/B receptor antagonists in early diabetic kidney disease. J Pharmacol Exp Ther. 2011;338(1):263‐270. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Sasser JM, Sullivan JC, Hobbs JL, et al. Endothelin a receptor blockade reduces diabetic renal injury via an anti‐inflammatory mechanism. J Am Soc Nephrol. 2007;18(1):143‐154. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Simonson MS, Ismail‐Beigi F. Endothelin‐1 increases collagen accumulation in renal mesangial cells by stimulating a chemokine and cytokine autocrine signaling loop. J Biol Chem. 2011;286(13):11003‐11008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Zoja C, Cattaneo S, Fiordaliso F, et al. Distinct cardiac and renal effects of ETA receptor antagonist and ACE inhibitor in experimental type 2 diabetes. Am J Physiol Renal Physiol. 2011;301(5):F1114‐F1123. [DOI] [PubMed] [Google Scholar]
- 21. Ruster C, Wolf G. Angiotensin II as a morphogenic cytokine stimulating renal fibrogenesis. J Am Soc Nephrol. 2011;22(7):1189‐1199. [DOI] [PubMed] [Google Scholar]
- 22. Trachtman H, Nelson P, Adler S, et al. DUET: a phase 2 study evaluating the efficacy and safety of sparsentan in patients with FSGS. J Am Soc Nephrol. 2018;29(11):2745‐2754. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Komers R, Diva U, Inrig JK, Loewen A, Trachtman H, Rote WE. Study design of the phase 3 sparsentan versus irbesartan (DUPLEX) study in patients with focal segmental glomerulosclerosis. Kidney Int Rep. 2020;5(4):494‐502. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Troost JP, Trachtman H, Nachman PH, et al. An outcomes‐based definition of proteinuria remission in focal segmental glomerulosclerosis. Clin J Am Soc Nephrol. 2018;13(3):414‐421. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Yafune A, Ishiguro M. Bootstrap approach for constructing confidence intervals for population pharmacokinetic parameters. I: a use of bootstrap standard error. Stat Med. 1999;18(5):581‐599. [DOI] [PubMed] [Google Scholar]
- 26. Tran JQ, Kovacs SJ, McIntosh TS, Davis HM, Martin DE. Morning spot and 24‐hour urinary 6 beta‐hydroxycortisol to cortisol ratios: intraindividual variability and correlation under basal conditions and conditions of CYP 3A4 induction. J Clin Pharmacol. 1999;39(5):487‐494. [PubMed] [Google Scholar]
- 27. Volpe DA, Tobin GA, Tavakkoli F, Dowling TC, Light PD, Parker RJ. Effect of uremic serum and uremic toxins on drug metabolism in human microsomes. Regul Toxicol Pharmacol. 2014;68(2):297‐303. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure S1
Figure S2
Figure S3
Table S1
Table S2
Table S3
Table S4
Table S5
Data Availability Statement
The data used for the analyses in this manuscript are available on request from the corresponding author.