No QTcF Prolongation with Sepiapterin: Results From a Thorough QT Study in Healthy Subjects at Therapeutic and Supratherapeutic Doses

Abstract

Sepiapterin and its major metabolite 6R‐L‐erythro‐5,6,7,8‐tetrahydrobiopterin (BH₄) bind to distinct variants of phenylalanine hydroxylase (PAH), which converts excess phenylalanine to tyrosine, thereby stabilizing, enhancing, and prolonging PAH activity. Sepiapterin was recently approved in Europe and the USA for the treatment of hyperphenylalaninemia patients with phenylketonuria, an inherent metabolic disease caused by PAH deficiency. A thorough QT study of sepiapterin in healthy volunteers at therapeutic (60 mg/kg) and supratherapeutic (120 mg/kg) doses was conducted to assess potential cardiovascular risks. Thirty‐two participants were randomized into one of 12 sequences and received single doses of sepiapterin (60 or 120 mg/kg), moxifloxacin 400 mg, or placebo in separate periods. Sepiapterin had no effect on heart rate or cardiac conduction (PR/QRS interval). Saturable sepiapterin absorption was observed, which resulted in less than dose‐proportional increase of sepiapterin and BH₄ and limited the maximum plasma concentrations clinically achievable. Using concentration‐QT analysis, the placebo‐corrected change from baseline in QT interval corrected using Fridericia's formula (ΔΔQTcF) was −2.11 (90% CI: −3.44, −0.79) ms at geometric mean baseline‐corrected BH₄ C_max (728 ng/mL) and −1.9 (−3.25, −0.56) ms at sepiapterin C_max (2.08 ng/mL) at the supratherapeutic dose of 120 mg/kg. An effect on ΔΔQTcF exceeding 10 ms was excluded within the observed concentration range of baseline‐corrected BH₄ up to 1088 ng/mL and sepiapterin up to 5.77 ng/mL. The consistency of results from this study and the previous concentration‐QTc analysis based on pooled data from multiple clinical studies demonstrated the reliability of using concentration‐QTc for assessing cardiovascular risks in early clinical development.

Introduction

Phenylketonuria (PKU) is an autosomal, recessive, genetic disorder, first identified in 1934 by Følling when he detected phenylketone bodies in urine. ¹ Since the introduction of the Guthrie dried blood method in the 1960s, PKU is now diagnosed at birth through newborn screening worldwide. PKU is caused by deficiency of the enzyme phenylalanine hydroxylase (PAH), which converts phenylalanine (Phe) to tyrosine. Deficiency of PAH leads to increased phenylalanine in the blood and brain. Untreated, PKU can lead to irreversible intellectual disability, microcephaly, motor deficits, and developmental problems. Current PKU treatment guidelines aim to decrease blood Phe concentrations. ¹

Sepiapterin (marketed under the name Sephience) is a new molecular entity, an exogenously synthesized, structurally equivalent molecule of biologically produced sepiapterin. Sepiapterin functions in a dual capacity as (1) an endogenous, naturally occurring precursor to 6R‐L‐erythro‐5,6,7,8‐tetrahydrobiopterin (BH₄) via the pterin salvage pathway ² and (2) a pharmacologically active chaperone for the native state conformation of PAH, conferring stability against thermal unfolding, resulting in enhanced activity and prolongation of PAH function. ³ The efficacy of sepiapterin for the treatment of adult and pediatric patients with PKU has been demonstrated in multiple clinical studies. ⁴ , ⁵ , ⁶ Sepiapterin was recently approved for the treatment of adult and pediatric patients with PKU in Europe (June 2025), the USA (July 2025), and multiple countries worldwide.

Following oral administration, sepiapterin is quickly absorbed, rapidly and extensively converted to BH₄ exclusively intracellularly. ² , ⁷ , ⁸ Peak plasma concentration (C_max) of sepiapterin is reached approximately 1–3 h (T_max) after administration and declines quickly to below the lower limit of quantification (LLOQ, 0.75 ng/mL) by 12 h. The C_max of BH₄ is reached at approximately 4 h (T_max) and thereafter decreases steadily with an apparent terminal elimination half‐life (T_1/2) of 5 h. ⁸ No accumulation of BH₄ was observed in healthy volunteers at sepiapterin doses up to 60 mg/kg once daily administered with a high‐fat diet. ² The sepiapterin C_max and area under the plasma drug concentration‐time curve (AUC) were generally <2% of the values of BH₄ following oral administration of sepiapterin. Administration of sepiapterin with food increased BH₄ C_max and AUC, and a greater increase was observed with a high‐fat diet than with a low‐fat diet. ² , ⁸ Both BH₄ C_max and AUC increased with escalating doses of sepiapterin, but the increase was far less than dose proportional when the sepiapterin dose was above 20 mg/kg. ⁸

Pharmacokinetic (PK) blood samples and time‐matched electrocardiogram (ECG) data have been collected from healthy adults in several phase 1 studies ⁷ , ⁸ , ⁹ and from adult and pediatric patients with PKU in the pivotal phase 3 study. ⁴ A concentration‐QTc (C‐QT) analysis conducted from the pooled data demonstrated that there were no QT prolongation risks associated with sepiapterin at therapeutic doses up to 60 mg/kg; ¹⁰ however, this analysis had some limitations: (1) no data were available from participants treated with placebo; and (2) it lacked clinical data at a supratherapeutic sepiapterin dose (the highest dose used was the therapeutic dose of 60 mg/kg). The sensitivity of this analysis was demonstrated through food effects on QT interval shortening. The mean BH₄ C_max observed in healthy adults was 1.97‐fold greater than that observed in PKU patients, even though they received the same sepiapterin dose, which was attributed to food effects. In healthy adults, sepiapterin was administered with a high‐fat diet, while it was administered with an individualized Phe‐restricted diet for PKU patients—the fat content in the Phe‐restricted diet was generally similar to a standard low‐fat diet. ¹¹

A phase 1, randomized, partial‐blinded, placebo‐ and positive‐controlled, 4‐way crossover thorough QT (TQT) study was performed to evaluate the potential cardiovascular effect of sepiapterin. Healthy participants received a single dose of sepiapterin at either the therapeutic dose of 60 mg/kg or the supratherapeutic dose of 120 mg/kg. The supratherapeutic dose 120 mg/kg selected was based on the gastrointestinal tolerability and potential maximum exposures that could be achieved. Previous clinical data have already demonstrated that the increase of BH₄ exposure is far less than dose proportional at sepiapterin doses above 20 mg/kg; hence, it is expected that limited increases of BH₄ exposure can be achieved at doses above 120 mg/kg. Clinical studies have shown that administering sepiapterin with a low‐fat diet increases BH₄ C_max and AUC from time zero to 24 h after drug administration (AUC_0–24 h) 1.7‐fold and 1.7‐fold, respectively, and a high‐fat diet increases BH₄ C_max and AUC_0–24h 2.2‐fold and 2.8‐fold, respectively. ⁸ Thus, all treatments were administered with a high‐fat diet in the study to increase the C_max that could be reached. Moxifloxacin (400 mg orally) treatment was the positive control in the study to demonstrate sensitivity.

Methods

Informed Consent and Ethics

The study was conducted in accordance with the International Council on Harmonisation ethical principles of Good Clinical Practice and the Declaration of Helsinki. The final protocol and informed consent form were reviewed and approved by Advarra Institutional Review Board (Columbia, MD, USA) and a written informed consent form was obtained from each participant.

Clinical Study Design

This was a phase 1, randomized, partially blinded (moxifloxacin treatment was not blinded), placebo‐ and positive‐controlled, 4‐way crossover study conducted in healthy adults at a single center in the USA (ICON Early Development Services, Lenexa, KS, USA). Eligible participants were randomized into one of 12 sequences, each consisting of four single‐dose treatment periods (sepiapterin 60 mg/kg, sepiapterin 120 mg/kg, placebo, or moxifloxacin 400 mg) (Figure S1). There was a minimum washout period of 4 days between treatments. All treatments were administered with a high‐fat, high‐calorie meal after a minimum of 10 h overnight fasting. All study treatments were administered with 240 mL of water within 30 min after starting the high‐fat, high‐calorie meal, which was described in the FDA guidance, ¹² and within 5 min after consumption of the meal. Participants were restricted from food consumption within the first 4 h after treatment administration. Fluid intake was restricted from 1 h before dosing to 2 h after dosing, except for those served as part of a meal before dosing and the water (240 mL) used for dose administration. Caffeine, alcohol, and tobacco consumption were prohibited for the duration of the study.

Study Participants

The study included males and females aged 18–55 years in good health with a body mass index (BMI) of 18–33 kg/m², body weight of 50–100 kg, no chronic medical conditions, and no clinically significant physical exam or laboratory findings. Subjects were excluded from the study if any of the following applied: QRS interval > 120 ms; QTc using Fridericia's formula (QTcF) > 450 ms (males) or > 470 ms (females); PR interval > 220 ms; history of risk factors for torsades de pointes; sustained supine systolic blood pressure > 150 mmHg or < 90 mmHg, or a supine diastolic blood pressure > 95 mmHg or < 50 mmHg; a resting heart rate (HR) of < 40 bpm or > 100 bpm; and a glomerular filtration rate (GFR) < 80 mL/min.

Blood Sample Collection and Concentration Measurement

Blood samples were collected in each period at −75, −60, and −45 min before dosing, and 0.5, 1, 2, 3, 4, 5, 6, 8, 10, 12, 24, and 32 h after dosing into tubes prefilled with dipotassium ethylenediamin tetra‐acetic acid (K₂EDTA) as an anticoagulant and put in wet ice. Blood samples collected following sepiapterin or placebo treatment were mixed with 10% ascorbic acid (prepared within 1 week before use) in a 10:1 (v:v) ratio within 15 min of blood collection to stabilize sepiapterin and BH_4. ¹³ All blood samples were centrifuged at 2000× g and 4°C for 10 min within 30 min of collection to harvest plasma samples. Plasma samples were frozen within 30 min of blood collection at −70°C until sample analysis and shipped on dry ice. Validated high‐performance liquid chromatography with tandem mass spectrometry methods were used to measure plasma concentrations of sepiapterin (LLOQ 0.75 ng/mL), BH₄ (LLOQ 0.5 ng/mL), and moxifloxacin (LLOQ 10.0 ng/mL). ¹³

ECG Measurements

Continuous ECG recordings (Holters) were performed starting 75 min before dosing on day 1 until the last PK sample collection and safety ECG measurement in each period. Twelve‐lead ECGs were extracted in up to 10 replicates from the continuous Holter recording using the Early Precision QT (EPQT) method by a central ECG Laboratory (eResearch Technologies, Inc. [ERT], a Clario Company, Philadelphia, PA, USA) at the following time points: three pre‐dose time points (−75, −60, and −45 min), and at 0.5, 1, 2, 3, 4, 5, 6, 8, 10, 12, 24, and 32 h after dosing, for a total of 15 time points. Participants were instructed to rest in a supine position for at least 15 min before the nominal time point. ECGs were extracted within a window of approximately 5 min (at time points ≤ 8 h after dosing) or 10 min (at time points > 8 h after dosing) before time‐matched PK sample collection by the central ECG laboratory. A total of ten 14‐s, digital, 12‐lead ECG tracings were extracted from the continuous Holter recordings using the TQT Plus method, ¹⁴ , ¹⁵ which used computer derived QT measurements while maintaining human oversight of quality and of scoring of T/U wave morphology changes. Compared with triplicated ECG reading using semi‐automated methods, using this EPQT method, the standard deviation of change‐from‐baseline QTc is substantially lower, and the sample size can be further reduced while maintaining the power. ¹⁵ , ¹⁶

PK Analysis

PK parameters of sepiapterin, BH₄, and moxifloxacin were derived using Phoenix WinNonlin version 8.0 or higher (Certara, Princeton, NJ, USA) with the non‐compartmental method based on actual sampling times. AUCs were calculated following the linear up‐logarithm down algorithm. Both sepiapterin and BH₄ are endogenous compounds; however, endogenous sepiapterin is always below the LLOQ (BLQ, LLOQ 0.75 ng/mL), while endogenous BH₄ is measurable (LLOQ 0.5 ng/mL). ⁸ Hence, baseline BH₄ was defined for each individual participant for every sepiapterin treatment period as the arithmetic mean of pre‐dose measurements. Baseline‐corrected BH₄ concentrations were derived for each participant at every time point by subtracting the baseline BH₄ concentration. The baseline‐corrected BH₄ concentrations were assigned as zero if a negative value was generated from the subtraction. PK parameters for BH₄ were calculated using both uncorrected BH₄ concentrations and baseline‐corrected concentrations. Since the baseline BH₄ concentrations were negligible compared with those raised after sepiapterin treatment, only PK parameters derived from uncorrected BH₄ were presented in this article to avoid redundancy and for easier comparison with the literature. BLQ concentrations for sepiapterin and moxifloxacin were treated as zero. As the geometric mean C_max of each analyte was used to assess potential QT prolongations, all PK parameters in this study were presented as geometric mean (percent of geometric coefficient of variation, GCV%).

PK Dose Linearity Analysis

In this study, 2 sepiapterin dose levels (60 and 120 mg/kg) were studied. To assess dose linearity over a broad dose range, exposures to BH₄ (C_max and AUC_0–24 h) from this study at sepiapterin doses of 60 and 120 mg/kg were pooled together with data obtained from healthy volunteers from previous clinical studies ⁷ , ⁸ at various sepiapterin doses administered with high‐fat meals for the analysis. Both the linear mixed‐effect power model with logarithm transformed PK parameters ¹⁷ and the saturable maximum effect (E_max) model, described by Equation (1) below, were explored to describe the relationship:

(1)

where E_max and K_m represent the maximum value and dose at which 50% of the maximum value for a given PK parameter could be observed and dose is the sepiapterin dose expressed in mg/kg.

Dose linearity was not assessed for sepiapterin because sepiapterin concentrations were below 2% of BH₄, and concentrations were below the LLOQ at many time points.

By‐Time‐Point QT Analysis

The by‐time‐point analysis of QTcF values was performed using a mixed model for repeated measures, with change in QTcF from baseline (ΔQTcF) as the dependent variable; period, sequence, time (i.e., nominal post‐dose time point), treatment (sepiapterin 60 mg/kg, sepiapterin 120 mg/kg, moxifloxacin, placebo), and time‐by‐treatment interaction as fixed effects, and baseline QTc as a covariate. An unstructured covariance matrix was specified for the repeated measures at post‐dose time points for participants during the treatment period. From this analysis, the least‐squares (LS) mean, standard error (SE), and two‐sided 90% confidence interval (CI) of placebo‐corrected ΔQTcF (ΔΔQTcF) were calculated for the contrast “sepiapterin versus placebo” for each dose of sepiapterin and for the contrast “moxifloxacin versus placebo” for moxifloxacin at each post‐dose time point separately. For heart rate (HR), PR interval, and QRS interval, the analysis was based on the change from baseline at post‐dose values (ΔHR, ΔPR, ΔQRS). The same (by‐time‐point analysis) model was used as described for the QTcF interval.

C‐QT Analysis

Before the C‐QT analysis was performed, the following model‐independent assumptions were explored graphically: there was no drug effect on HR; QT from the selected primary QT correction method (Fridericia method) was independent of HR; and there was no apparent time delay between drug concentrations and ΔΔQTcF (hysteresis), and a linear C‐QT relationship.

The relationship between ΔQTcF and plasma concentrations of sepiapterin, BH₄, and moxifloxacin was investigated using the linear mixed‐effects modeling approach. For BH₄, baseline‐corrected BH₄ concentrations were used to correct the endogenous BH₄ effect. Study treatment (active = 1 or placebo = 0: continuous) and time (i.e., post‐dose time point: categorical) were explored as fixed effects and random effects on intercept and slope per participant. A population mean intercept was estimated. The centered baseline QTcF (i.e., baseline QTcF for an individual participant minus the population mean baseline QTcF for all participants in the same treatment period) was investigated as an additional covariate. From the model, the slope(s) and the treatment‐effect‐specific intercept (defined as the difference of “treatment” estimation between active drug and placebo) were estimated, together with the 2‐sided 90% CI. The point estimate and 2‐sided 90% CI for ΔΔQTcF (i.e., slope estimate × [geometric mean C_max of the analyte − geometric mean C_max of placebo (applicable if the post‐dose concentrations were not set to zero)] + treatment‐effect‐specific intercept) was obtained. If the upper bound of the 2‐sided 90% CI (equivalent to the upper bound of the one‐sided 95% CI) was below 10 ms, this would demonstrate that there is no clinically concerning QTc prolongation within the observed plasma concentration ranges.

Since sepiapterin is a naturally occurring precursor of BH₄, a similar analysis was performed for sepiapterin concentrations. An analysis including both BH₄ and sepiapterin was to be performed only when the separate models demonstrated that both BH₄ and sepiapterin had concentration‐dependent QT prolongation effects, and a model selection procedure with the Akaike information criterion (AIC) would be used to select the final model.

Sample Size

A sample size of 32 participants was planned to ensure that at least 24 evaluable participants would complete the study. A sample size of 24 evaluable participants provided more than 90% power to exclude the prediction that sepiapterin causes at least 10 ms QTc effect at clinically relevant plasma levels, as shown by the upper bound of the two‐sided 90% CI of the model‐estimated QTc effect (∆∆QTcF) at the observed geometric mean C_max of sepiapterin and BH₄ in the study. This power was estimated using a paired‐sample t‐test. The calculation assumed a onesided 5% significance level, a small underlying effect of sepiapterin of 3 ms, and a standard deviation (SD) of ΔΔQTcF of 8 $\sqrt{2}$  ms for both doses of sepiapterin. To demonstrate assay sensitivity with C‐QTc analysis, the ΔΔQTcF in response to a single dose of moxifloxacin 400 mg was expected to be more than 5 ms (i.e., the lower bound of the two‐sided 90% CI of the estimated QTc effect [ΔΔQTcF] should exceed 5 ms) at the observed geometric mean of C_max of moxifloxacin 400 mg. Under the assumption that the moxifloxacin QTc effect was 10 ms with an SD of ΔΔQTcF of 8 $\sqrt{2}$ ms for moxifloxacin, a sample size of 24 evaluable participants would provide 68% power to demonstrate assay sensitivity at a one‐sided 5% significance level. The proposed number of participants also aligned with recent FDA recommendations to enroll at least 20 participants. ¹⁸ Efforts were made to ensure that each sex was represented by at least 30% of participants.

Results

Study Population

In total, 32 healthy adults participated in the study, 31 of whom completed all four periods of treatment. One participant dropped out after receiving a placebo during the first treatment period due to personal reasons, which were not related to any adverse events. Baseline demographics of the 32 participants are summarized in Table 1. The mean (minimum, maximum) age was 38.4 (24, 54) years with equal numbers of males (n  =  16) and females (n  =  16). Mean body weight was 76.9 (54.3, 97.2) kg and mean BMI was 26.2 (20.7, 32.3) kg/m². The majority of participants were White (71.9%) and not Hispanic or Latino (71.9%).

Table 1.

Participants Demographic Characteristics

Category	Values (N = 32)
Age, years
Mean (min, max)	38.4 (24, 54)
Sex, n (%)
Female	16 (50.0)
Male	16 (50.0)
Race, n (%)
Asian	1 (3.1)
Black or African American	7 (21.9)
White	23 (71.9)
Multiple	1 (3.1)
Ethnicity, n (%)
Hispanic or Latino	9 (28.1)
Not Hispanic or Latino	23 (71.9)
Body weight, kg
Mean (min, max)	76.9 (54.3, 97.2)
BMI, kg/m2
Mean (min, max)	26.2 (20.7, 32.3)

BMI, body mass index; max, maximum; min, minimum.

Safety Results

There were three treatment‐emergent adverse events (TEAEs) reported for each treatment. The TEAEs reported during placebo treatment period included diarrhea, dyspepsia, and dermatitis contact (one of each). Abdominal discomfort, vessel puncture site injury, and genital lesion (one of each) reported during sepiapterin 60 mg/kg treatment period, while constipation, contusion, and dizziness (one of each) were reported during sepiapterin 120 mg/kg treatment period. During the moxifloxacin 400 mg treatment period, constipation, vessel puncture site injury, and gastroenteritis (one of each) were reported. All TEAEs were mild in severity. There were no severe AEs (SAEs) and no TEAEs that led to discontinuation from the study. Most TEAEs were related to gastrointestinal discomfort.

PK Results

The geometric mean plasma concentrations of sepiapterin, BH₄, and moxifloxacin following sepiapterin, moxifloxacin, and placebo treatments are presented in Figure 1. Following oral administration, sepiapterin was quickly absorbed and reached sepiapterin C_max at approximately 4–5 h (T_max) (Table 2). Absorbed sepiapterin was rapidly and extensively converted to the active metabolite BH₄. Sepiapterin concentrations dropped to BLQ generally by 12 h after dosing (Figure 1). The geometric mean sepiapterin C_max (1.74 and 1.98 ng/mL at a dose of 60 mg/kg and 120 mg/kg, respectively) and AUC_0–24 h were <1% of those of BH₄ (Table 2). Because of limited time points with measurable sepiapterin concentration post T_max, the T_1/2 and AUC_0‐inf could not be reliably estimated for sepiapterin.

Figure 1.

Geometric mean plasma concentrations of sepiapterin, BH₄, and moxifloxacin following a single oral dose of sepiapterin 60 or 120 mg/kg, placebo, or moxifloxacin 400 mg. Data were nudged along x‐axis for better visual presentation.

Table 2.

Geometric Mean (GCV%) of Pharmacokinetic Parameters in Healthy Adults (n = 31) Following a Single Oral Dose of Sepiapterin 60 mg/kg, Sepiapterin 120 mg/kg, or Moxifloxacin 400 mg With a High‐Fat, High‐Calorie Diet

Treatment	Analyte	Tmax a (h)	Cmax (ng/mL)	AUC0–24h (h·ng/mL)	AUC0–32h (h·ng/mL)	AUC0–inf (h·ng/mL)	T1/2 (h)
Sepiapterin 60 mg/kg	Sepiapterin b	4.00 (1.00, 10.00)	1.74 (43.2%)	8.83 (73.1%)	8.83 (73.1%)	NA	NA
	BH4	5.00 (4.00, 8.00)	640 (25.8%)	4909 (26.9%)	5008 (26.7%)	5070 (26.3%)	4.31 (23.2%)
Sepiapterin 120 mg/kg	Sepiapterin	5.00 (1.00, 10.13)	1.98 (51.8%)	14.3 (80.1%)	14.3 (80.1%)	NA	NA
	BH4	5.00 (4.00, 10.13)	738 (22.1%)	6246 (28.7%)	6388 (28.4%)	6556 (27.6%)	4.17 (22.3%)
Moxifloxacin 400 mg	Moxifloxacin	3.00 (0.50, 8.00)	1585 (25.3%)	19,993 (21.1%)	22,566 (20.7%)	26,406 (21.4%)	10.2 (14.4%)

NA, not applicable.

^a T_max was presented as median (minimum, maximum).

^b n = 30

Endogenous plasma BH₄ concentrations were measurable in samples from placebo treatment and pre‐dose samples before sepiapterin treatments. The endogenous BH₄ concentrations were log‐distributed as evidenced by the higher‐than‐center geometric mean concentration (Figure S2). The geometric mean (GCV%) of endogenous BH₄ concentrations was 2.37 ng/mL (26.4%) and the minimum and maximum values were 0.968 and 4.81 ng/mL. These values were far less than BH₄ C_max observed following sepiapterin treatments and are negligible for estimating BH₄ PK parameters (Figure 1 and Table S1). The mean BH₄ level from the three pre‐dose samples before sepiapterin treatment was treated as the baseline BH₄ level for each individual for every treatment period and was utilized to derive baseline‐corrected BH₄ concentrations.

Following sepiapterin administration, peak BH₄ concentrations were observed at 5 h (Table 2) and declined monophasically thereafter. The BH₄ C_max observed following dosing of sepiapterin 60 and 120 mg/kg was 640 and 738 ng/mL, respectively. The terminal T_1/2 was 4.31 and 4.17 h, respectively.

Moxifloxacin C_max was reached 3 h after dosing, with a geometric mean C_max of 1585 ng/mL (Table 2).

PK Dose Linearity

Sepiapterin C_max and AUC were insensitive to sepiapterin doses above 20 ng/mL, and there was large variability due to low concentrations, which were only measurable at limited time points. Hence, the dose linearity for the phase 3 formulation administered with a high‐fat diet was assessed with BH₄ C_max and AUC_0–24 h only. As sepiapterin dose increased from 60 to 120 mg/kg, a two‐fold dose increase, BH₄ C_max increased from 640 to 738 ng/mL and AUC_0–24 h increased from 4909 to 6246 h ng/mL (Table 2), an increase of 1.15‐ and 1.27‐fold, respectively. The results indicated that BH₄ exposures increased far less than dose proportionally with sepiapterin 60–120 mg/kg.

An analysis based on pooled data from this study and earlier clinical studies ⁷ , ⁸ in healthy adults who received sepiapterin with a high‐fat diet demonstrated that both BH₄ C_max and AUC_0–24h were saturable (Figure 2). The dose‐exposure relationship could be described with an E_max equation (Equation (1)). The E_max and K_m were 820.1 ng/mL and 17.19 mg/kg for BH₄ C_max and 9059 h·ng/mL and 43.68 mg/kg for BH₄ AUC_0–24h (Table 3). The Hill coefficients were 1.1857 and 0.8924 for BH₄ C_max and AUC_0–24h, respectively, both close to the unity value 1. A further analysis after splitting the dose range into 5–20 mg/kg and 20–120 mg/kg indicated that BH₄ C_max and AUC_0–24 h increased approximately dose proportionally in the dose range 5–20 mg/kg (the slopes were within 0.80–1.25 following the power model), but less than dose proportionally at doses above 20 mg/kg (Table S2).

Figure 2.

Relationship of BH₄ (a) C_max and (b) AUC_0–24 h with sepiapterin dose. Open red circle, arithmetic mean of observed data at each dose; black error bar, standard deviation at each dose; dashed blue line, regression line from an E_max equation fitting.

Table 3.

Regression of BH₄ C_max and AUC_0–24 h with Sepiapterin Dose Administered with a High‐Fat, High‐Calorie Diet Using an E_max Equation from Pooled Data in Healthy Adults

Parameter		Emax (ng/mL or h·ng/mL)	Km (mg/kg)	Hill
Cmax	Estimate	820.07	17.19	1.1857
	90% CI	696.8–943.4	11.14–23.25	0.7555–1.6160
AUC0–24 h	Estimate	9059	43.68	0.8924
	90% CI	5297–12820	−2.00–89.35	0.5154–1.2695

E_max, maximum effect; K_m, dose at which 50% of the maximum value for a given pharmacokinetic parameter could be observed.

Heart Rate and Cardiac Conduction Results

Sepiapterin did not exhibit any clinically relevant effects on HR at the studied doses. The LS mean placebo‐corrected change from baseline in HR (ΔΔHR) across post‐dose time points varied from −0.5 bpm (beat per minute) (at 32 h after dosing) to 1.2 bpm (at 4 h after dosing) with sepiapterin 60 mg/kg, and from −1.6 bpm (32 h after dosing) to 2.5 bpm (8 h after dosing) with sepiapterin 120 mg/kg. The graph of QTcF versus reciprocal of the HR (RR) indicated that QTcF was independent of the RR and that QTcF was thereby an appropriate method for QT correction (QTc) (Figure S3).

Sepiapterin did not have any clinically relevant effect on cardiac conduction (i.e., the PR and QRS intervals). LS mean placebo‐corrected change from baseline in PR interval (ΔΔPR) across post‐dose time points varied between −2.1 ms (3 h and 8 h after dosing) and 1.7 ms (0.5 h after dosing) with sepiapterin 60 mg/kg and between −2.7 ms (8 h after dosing) and 1.6 ms (24 h after dosing) with sepiapterin 120 mg/kg (Figure S4a). LS mean placebo‐corrected change from baseline in QRS interval (ΔΔQRS) was within ± 1.0 ms across all post‐dose time points (Figure S4b).

By‐Time‐Point QT Analysis

LS mean ΔQTcF following sepiapterin oral administration generally followed the placebo pattern across post‐dose time points (Figure 3a). LS mean ΔΔQTcF across post‐dose time points ranged from −3.2 ms (6 h after dosing) in the 60 mg/kg dose group to 3.0 ms (0.5 h after dosing) in the 120 mg/kg dose group, without dose‐dependency. The upper limit of the 90% CI of ΔΔQTcF in the by‐time‐point analysis was <10 ms at sepiapterin 60 and 120 mg/kg across all time points, while it was >5 ms for moxifloxacin between 3 and 12 h after dosing (Figure 3b).

Figure 3.

LS mean and 90%CI of (a) ΔQTcF and (b) ΔΔQTcF at each time point following treatment with sepiapterin 60 mg/kg, sepiapterin 120 mg/kg, moxifloxacin 400 mg, and placebo.

Hysteresis Analysis

A key prerequisite for the applicability of model‐based C‐QTc analysis is the absence of hysteresis, a delay between the effect of drug on the QTcF (ΔΔQTcF) and plasma concentrations. The hysteresis of QTcF with sepiapterin and BH₄ concentrations following sepiapterin doses was assessed graphically. The time courses of geometric mean sepiapterin plasma concentrations and LS mean ΔΔQTcF across post‐dose time points for sepiapterin 60 and 120 mg/kg are illustrated in Figure S5a,b. Similarly, the time courses of geometric mean baseline‐corrected BH₄ plasma concentrations and LS mean ΔΔQTcF across post‐dose time points for sepiapterin 60 and 120 mg/kg are illustrated in Figure S5c,d. There were inconsistencies in the difference (delay) between the time to reach the largest QT effect (i.e., ΔΔQTcF) and geometric mean plasma concentrations of sepiapterin or baseline‐corrected BH₄, and there lacked a meaningful change in the ΔΔQTcF. Together, it indicated that the hysteresis of the QT effect was not present.

Model‐Based C‐QT Analysis

The relationship between ΔQTcF and individual observed baseline‐corrected BH₄ concentrations is shown in Figure 4a. The linear regression line was essentially superimposed with the locally estimated scatterplot smoothing (LOESS) across all observed baseline‐corrected BH₄ concentration ranges, and the line diverged across the concentration range. Additionally, sepiapterin did not have any clinically relevant effects on HR or cardiac conduction (PR and QRS intervals), QTcF was independent of RR, and there was no hysteresis of QT effect with either sepiapterin or baseline‐corrected BH₄ concentrations. Hence, it was appropriate to utilize linear models for C‐QT relationship to assess sepiapterin treatment effects on the QTc interval.

Figure 4.

Relationship of (a) observed ΔQTcF with baseline‐corrected BH₄ concentrations and (b) model‐estimated ΔΔQTcF (mean and 90% CI) and deciles of baseline‐corrected BH₄ concentrations at sepiapterin dose 60 and 120 mg/kg. In (b), solid black line with gray shaded area: model‐estimated mean ΔΔQTcF and 90% CI; red filled circle and error bar: estimated ΔΔQTcF with 90% CI at associated median concentration within each decile; horizontal red line with notches: range of concentrations divided into deciles; horizontal dashed gray line: 10 ms threshold.

The relationship between model‐estimated ΔΔQTcF and deciles of baseline‐corrected BH₄ concentrations (Figure 4b) illustrated the adequacy of the linear mixed‐effects model for the C‐QT analysis. The mean ΔΔQTcF at the associated median plasma concentration within each decile of baseline‐corrected BH₄ concentrations was within the 90% CI of model‐estimated mean ΔΔQTcF, except for the third decile. Thus, it could be concluded that the linear mixed‐effect C‐QT model adequately described the relationship between ΔΔQTcF and baseline‐corrected BH₄ concentrations.

The estimated population slope of the C‐QT relationship with baseline‐corrected BH₄ concentrations was −0.0028 ms per ng/mL (90% CI:−0.00518, −0.00052; P = 0.0449) with a very small treatment‐effect‐specific intercept of −0.046 ms (90% CI: −0.9139, 0.8225; P = 0.9310). The slope of baseline‐corrected BH₄ plasma concentrations was statistically significant at the 10% significance level; the treatment‐effect‐specific intercept was not statistically significant. Based on the model, the estimated ΔΔQTcF at the geometric mean baseline‐corrected BH₄ C_max of 632 ng/mL (at dose 60 mg/kg) and 726 ng/mL (at dose 120 mg/kg) were −1.85 ms (90% CI: −2.99, −0.70 ms) and −2.11 ms (90% CI: −3.44, −0.79 ms). An effect on ΔΔQTcF exceeding 10 ms can be excluded within the full observed range of baseline‐corrected BH₄ plasma concentrations up to 1088 ng/mL (Table 4).

Table 4.

Model Estimated ΔΔQTcF at Observed Maximum Geometric Mean Maximum Plasma Concentrations Based on Linear Mixed‐Effect Models

Analyte (explanatory variable)	Treatment	Cmax, ng/mL	ΔΔQTcF estimate, ms (90% CI)
Baseline‐corrected BH4	Sepiapterin 60 mg/kg	632	−1.85 (−2.99, −0.70)
	Sepiapterin 120 mg/kg	726	−2.11 (−3.44, −0.79)
Sepiapterin	Sepiapterin 60 mg/kg	1.67	−1.60 (−2.70, −0.51)
	Sepiapterin 120 mg/kg	1.94	−1.80 (−3.05, −0.55)
Moxifloxacin	Moxifloxacin 400 mg	1585	10.01 (8.35, 11.67)

Note: ΔQTcF was the dependent variable, time‐matched baseline‐corrected BH₄ concentrations, sepiapterin concentrations, or moxifloxacin concentrations were explanatory variables, centered baseline QTcF was a covariate, treatment (active = 1 and placebo = 0) and time were fixed effects, with a random intercept and random slope per participant.

Similar analysis was conducted to assess the C‐QT relationship with sepiapterin concentrations (Figure S6). Results demonstrated that an effect on ΔΔQTcF exceeding 10 ms can be excluded within the full observed range of sepiapterin plasma concentrations up to 5.77 ng/mL (Table 4).

Assay Sensitivity

When moxifloxacin 400 mg was administered with a high‐fat, high‐calorie diet, plasma moxifloxacin C_max was observed 3 h after dosing (Table 2) and the maximum ΔΔQTcF of 10.5 ms (90% CI: 8.63, 12.42 ms) was observed at 4 h after dosing (Figure 3b).

The linear regression line and LOESS regression line of observed ΔQTcF and moxifloxacin concentrations were in close approximation (Figure S7a). The mean ΔΔQTcF within each moxifloxacin concentration decile at the median concentration was within the 90% CI of the model‐estimated mean ΔΔQTcF, except for the 5 and 7 deciles (Figure S7b). It was consequently concluded that the model adequately described the relationship between ΔΔQTcF and moxifloxacin concentrations.

The estimated population slope of the moxifloxacin C‐QT relationship was 0.0040 ms per ng/mL (90% CI: 0.00269, 0.00541; P < 0.0001), with a treatment‐effect‐specific intercept of 3.59 ms (90% CI: 2.452, 4.733; P < 0.0001). Both the treatment‐effect‐specific intercept and the slope of moxifloxacin concentration were statistically significant at the 10% significance level (Table S5). Based on the model, the estimated ΔΔQTcF at the geometric moxifloxacin C_max was 10.01 ms (90% CI: 8.35, 11.67 ms) (Table 4). Since the slope of the C‐QTc relationship was statistically significant and the lower bound of the 90% CI of ΔΔQTcF was estimated to be 8.35 ms, exceeding the 5 ms criterion at the geometric C_max 1585 ng/mL, the assay sensitivity was demonstrated.

Discussion

A C‐QT analysis was conducted previously with pooled data from phase 1 studies in healthy adults and a phase 3 study in patients aged 1.4–61 years with PKU. ¹⁰ The highest sepiapterin dose included in the analysis was the maximum approved therapeutic dose of 60 mg/kg in both healthy adults and patients. A BH₄ C_max margin of 1.97‐fold was attained between healthy adults and PKU patients, which was attributed to food effects. For treatment of PKU patients, sepiapterin was administered with an individualized Phe‐restricted diet, which generally contained a similar amount of fat as a standard low‐fat diet, ¹¹ , ¹² while some healthy adults received sepiapterin with a high‐fat, high‐calorie diet. A previous clinical study demonstrated that BH₄ C_max and AUC increased commensurately with the fat content of the diet. ⁸ Results from that C‐QT analysis demonstrated that there was no effect on ΔΔQTcF exceeding 10 ms across observed baseline‐corrected BH₄ concentrations up to 1050 ng/mL and sepiapterin concentrations up to 9 ng/mL. ¹⁰

Following oral administration, sepiapterin is quickly absorbed, rapidly and extensively converted to the active metabolite BH₄. ² , ⁸ BH₄ serves as a cofactor for NOS and facilitates the generation of nitric oxide (NO). ¹⁹ The binding of BH₄ to NOS stabilizes the enzyme, enhances the substrate arginine binding, and maintains molecular oxygen “coupled” to NOS, thereby facilitating electron flow from arginine to molecular oxygen during the production of NO. ²⁰ At suboptimal concentration of BH₄, or at the reduced ratio of BH₄/BH₂ (7,8‐dihydrobiopterin, the oxidized metabolite of BH₄), NOS is destabilized, referred to as “uncoupled”, resulting in production of superoxide radical (O₂ ^•−) instead on NO. ²¹ , ²² NO activates soluble guanylyl cyclase (sGC) in vascular smooth muscle cells, increasing cyclic guanosine monophosphate (cGMP) production and resulting in vasodilation. ²³ A reduction in endothelial NOS and NO production is closely associated with endothelial dysfunction in hypertension. ²⁴

The vasodilatory effect of BH₄ makes it a potential therapeutic option for improving endothelial dysfunction for treating patients with systemic or pulmonary hypertension. However, exploratory clinical trials with oral BH₄ administration failed to demonstrate any significant efficacy. ²¹ Vasodilation may also increase QTc interval. Sublingual administration of glyceryl trinitrate 0.5 mg induced a prolongation of QTc 5.7 ms in healthy male subjects. ²⁵ Similarly, oral administration at the therapeutic doses of vardenafil (10 mg) and sildenafil (50 mg), both introduced vasodilation by inhibiting phosphodiesterase‐5 (PDE‐5) in healthy male adults (ΔΔQTcF prolongation of 8 and 6 ms, respectively). ²⁶ Although the previously conducted C‐QT analysis for sepiapterin demonstrated that there was a negligible negative central trend in QTcF change, the highest dose included in that analysis was the therapeutic dose. ¹⁰ Additionally, that analysis lacked a positive control, though the food effect was included as a sensitivity analysis. Accordingly, the results of the previous C‐QT analysis may not fully account for the potential risks associated with supratherapeutic doses of sepiapterin.

In this TQT study, the therapeutic and supratherapeutic doses of sepiapterin (60 and 120 mg/kg) were administered with a high‐fat diet to healthy adults. The dose linearity analysis based on pooled data from this study and previous clinical trials clearly demonstrates that sepiapterin absorption was saturable. This saturation is likely due to limited sepiapterin solubility. Sepiapterin solubility is determined to be less than 2 mg/mL in an aqueous solution with a pH range from 2.3 to 8.8 at 37°C, which translates to less than 7 mg/kg for a typical adult with a body weight of 70 kg and stomach fluid volume of 250 mL. The K_m for BH₄ C_max and AUC_0–24h following sepiapterin oral administration were 17.19 and 43.68 mg/kg, respectively. The same magnitude of K_m for C_max and predicted sepiapterin dose limited by solubility confirms the hypothesis. The higher K_m for BH₄ AUC_0–24h is attributed to the possibility of an extended duration of absorption.

Following the supratherapeutic sepiapterin dose 120 mg/kg, the geometric mean baseline‐corrected BH₄ C_max was 728 ng/mL, 2.26‐fold of BH₄ C_max in PKU patients at the maximum recommended therapeutic dose of 60 mg/kg. Results from both the C‐QT analysis for BH₄ and sepiapterin and the by‐time‐point analysis demonstrated that there was no effect on ΔΔQTcF exceeding 10 ms across observed baseline‐corrected BH₄ concentrations up to 1088 ng/mL and sepiapterin concentrations up to 5.77 ng/mL. The TQT study confirmed the results of previous C‐QT analyses from pooled data of healthy adults and PKU patients. ¹⁰ Consistent results from this dedicated TQT study and C‐QT analysis based on pooled data from phase 1 and 3 studies demonstrated that the C‐QT analysis based on phase 1 study data from early clinical development was a reliable approach to assess QT prolongation risks.

The vasodilatory effect of BH₄ is unlikely to exert any substantial cardiovascular effect following sepiapterin oral administration. The PR and QRS intervals were stable following sepiapterin treatment (Figure S4) and were comparable to those receiving placebo and moxifloxacin treatments. Contrary to typical observations of QTcF prolongation following treatment with vasodilators such as glyceryl trinitrate, vardenafil, and sildenafil, a negligible QTcF shortening of −2.11 ms was observed with a supratherapeutic sepiaterin dose of 120 mg/kg, which was less than the effect of a meal to 4 h postprandial (−5.78 ms). ²⁷ This QTcF shortening was consistent with the previous C‐QT analysis (−1.25 ms) at therapeutic sepiapterin dose 60 mg/kg in healthy volunteers with pooled data from multiple clinical trials. ¹⁰ A similar, but slightly larger QTc shortening (−8.32 ms) has been reported when BH₄ was administered at the supratherapeutic dose 100 mg/kg (5× of the maximum recommended dose for treatment of PKU patients). ²⁸ The mechanism of this QT shortening is not yet fully understood. However, since the magnitude is consistently less than that of postprandial, it is deemed that there is no cardiovascular risk associated with sepiapterin treatment.

Conclusions

The study demonstrated that the linear mixed‐effect model was appropriate for analyzing relationships between sepiapterin and baseline‐corrected BH₄ concentrations with ΔQTcF. The consistency of results from this TQT study and previous C‐QT analysis (based on pooled data from phase 1 studies in healthy adults and phase 3 studies in patients with PKU) demonstrates the reliability of C‐QT analysis from phase 1 studies in early clinical development.

Results from both the linear models C‐QT analysis and by‐time‐point analysis, demonstrated that there was no effect on ΔΔQTcF exceeding 10 ms at the therapeutic dose 60 mg/kg and supratherapeutic sepiapterin dose of 120 mg/kg in the full range of observed baseline‐corrected BH₄ plasma concentrations up to 1088 ng/mL or sepiapterin concentrations up to 5.77 ng/mL.

Author Contributions

Lan Gao, Ronald Kong, Neil Smith, and Lee Golden designed the study. Hongqi Xue and Borje Darpo conducted cardiovascular effect analysis. Lan Gao conducted pharmacokinetic analysis. Kimberley Ingalls monitored the study. Diksha Kaushik supervised the blood sample analysis. Hongqi Xue and Lan Gao provided the tables and figures. Lan Gao drafted the manuscript. All authors reviewed, edited, and approved the manuscript.

Conflicts of Interest

Lan Gao, Kimberly Ingalls, Diksha Kaushik, Neil Smith, Ronald Kong, and Lee Golden are employees of PTC Therapeutics and may own shares in PTC Therapeutics, Inc. Hongqi Xue is an employee of Clario. Borje Darpo is a consultant for Clario and owns shares and is eligible for stock options with Clario. Clario received research funding from PTC Therapeutics to conduct the study.

Funding

This study was funded by PTC Therapeutics, Inc.

Supporting information

Supporting information

Data Availability Statement

Data are available upon request.