Ethnic Bridging of Sepiapterin in Chinese and Korean Populations Based on Predictions From Genetic Polymorphism of Breast Cancer Resistance Protein

ABSTRACT

Ethnic differences are crucial when considering the efficacy, safety, and dose of pharmaceuticals across diverse populations. The International Council for Harmonization of Technical Requirements for Pharmaceuticals for Human Use (ICH) guideline E5 addresses the acceptability of extrapolating foreign clinical data taking ethnic factors into consideration. Sepiapterin has recently been approved for the treatment of hyperphenylalaninemia (HPA) in patients with phenylketonuria (PKU) in Europe, the USA, and multiple additional countries worldwide. To date, no clinical trials have been conducted in the Chinese or Korean populations. An ethnic sensitivity analysis identified that the breast cancer resistance protein (BCRP) c.421C>A variant was the primary factor leading to ethnic differences in BH₄ exposures. A correlation was established and validated between the frequency of BCRP c.421C>A variant in ethnic groups and the relative C _max and AUC_0–24h of sepiapterin major active metabolite 5,6,7,8‐tetrahydrobiopterin (BH₄). Based on this correlation, it was predicted that compared to White, the mean BH₄ C _max and AUC_0–24h were 1.16‐fold and 1.23‐fold higher, respectively, in Chinese subjects, and 1.12‐fold and 1.17‐fold higher, respectively, in Korean subjects. These findings, including the clinically insignificant differences in PK exposures, the comprehensive evidence of sepiapterin's efficacy and safety, the recognition of PKU as a rare disease and designation of sepiapterin as an orphan drug for treatment of PKU in EU, the USA, Japan, South Korea, and several other countries, and the urgent unmet medical need, collectively support that conducting an ethnic bridging study in Chinese and Korean populations is not warranted.

Study Highlights

• What is the current knowledge on the topic?

  • ○
    Sepiapterin has recently been approved in Europe and the USA for the treatment of hyperphenylalaninemia (HPA) in patients with phenylketonuria (PKU). No Chinese or Korean participants were included in the clinical trials of sepiapterin during development. The majority of participants were White.
• What question did this study address?

  • ○
    To determine ethnic differences in sepiapterin pharmacokinetics among Chinese, Korean, and Japanese populations compared with White populations.
• What does this study add to our knowledge?

  • ○
    The BCRP c.421C>A genetic polymorphism has been identified as the main factor contributing to ethnic differences in the pharmacokinetics of sepiapterin. A correlation formula based on the frequency of the BCRP c.421C>A variant in different ethnic groups was proposed to predict differences in sepiapterin pharmacokinetics among Chinese, Korean, and Japanese populations compared with the White population. Predictions were validated using clinical data obtained from Japanese and non‐Japanese healthy volunteers in an ethnic bridging study. The analysis predicted slightly higher C _max and AUC_0–24h values for 5,6,7,8‐tetrahydrobiopterin (BH₄), the major active metabolite of sepiapterin, in Chinese and Korean patients compared with White patients. This difference is not considered to be clinically significant, and the same sepiapterin dose is therefore proposed for Chinese and Korean patients with PKU.
• How might this change clinical pharmacology or translational science?

  • ○
    The approach to bridge the gap in ethnic data by using ethnic sensitivity analysis and modeling predictions based on genetic polymorphism of the primary transporter responsible for ethnic difference, and together with consideration of clinical safety and efficacy responses, to support a waiver of conducting ethnic bridging clinical trials demonstrates an alternative strategy in the absence of local clinical data in the targeted population.

1. Introduction

In 1998, the International Council for Harmonization of Technical Requirements for Pharmaceuticals for Human Use (ICH) (formerly known as International Conference on Harmonization) issued the Efficacy Guideline E5 on Ethnic Factors in the Acceptability of Foreign Clinical Data to facilitate the extrapolation of clinical data generated in one region for regulatory filing purposes in another region, aiming to reduce unnecessary clinical studies [1]. Subsequently, in 2003 and 2006, a Questions and Answers document and its revision regarding the implementation of ICH E5 were released to provide further clarification [2]. Before ICH E5, regulatory authorities frequently requested duplicated data due to concerns regarding ethnic differences in efficacy and safety [3]. ICH E5 has addressed this by providing a practical approach to assessing a drug's sensitivity to ethnic factors, including the intrinsic characteristics of the patients and extrinsic characteristics associated with environment and culture that could affect the results of clinical studies carried out in regions, and described the concept of the “bridging study” that a new region may use to determine whether data from another region are applicable to its population [1]. This approach enables the use of foreign clinical data to support registration approval in a new region, provided that an ethnic bridging study confirms the drug will behave similarly in both regions. Since the adoption of ICH E5 in Japan in 1998 [4], extensive research has demonstrated that the majority of approved drugs have similar exposures between Japanese and White populations [5, 6]. However, compared with the percentage of drugs with differences in pharmacokinetics (PK), a much higher percentage of approved drugs have a lower approved dose in Japan than in the USA or other Western countries [6]. The analysis also indicated that more drugs approved after the adoption of ICH E5 tend to have similar doses in both regions compared with those approved prior to the adoption of ICH E5 [6].

In 2018, the National Medical Products Administration (NMPA) in China issued a guidance on accepting data from foreign clinical trials [7]. In 2020, NMPA released a guidance on waiving the requirement for ethnic bridging studies for medicines already approved in foreign countries but not in China, based on the totality of efficacy, safety, and ethnic sensitivity [8]. The Pharmaceuticals and Medical Devices Agency (PMDA) in Japan published a guidance in 2023, which, in principle, waives the requirement for conducting phase I ethnic bridging studies before enrolling Japanese participants in multi‐regional clinical trials (MRCTs) [9, 10]. Although the 2023 PMDA's guideline only provides an opinion on enrolling Japanese participants in MRCTs, rather than waiving the requirement of ethnic bridging studies for new drug applications (NDA), it indicates a trend of shifting toward broader acceptance of multifactorial analysis in assessing ethnic differences in efficacy and safety data for drug approval.

Phenylketonuria (PKU) is an autosomal recessive inborn error of metabolism characterized by a deficiency of phenylalanine hydroxylase (PAH), which metabolizes phenylalanine (Phe) to form tyrosine [11, 12]. Gene mutations of PAH lead to decreased catalytic activity, causing hyperphenylalaninemia (HPA) [13], which, if left untreated, can lead to severe and irreversible intellectual disability [14]. PKU is diagnosed at birth in many developed countries due to the adoption of newborn screening. PKU has been identified in all ethnic groups, with an estimated global occurrence of approximately 1 in every 23,930 births [15]. The prevalence varies greatly between ethnic groups and regions, ranging from 1:125,000 in Japan to 1:4500 in Italy [15]. The prevalence is estimated to be 1:25,000 in the USA, 1:15,924 in China, and 1:41,000 in Korea [15]. PKU is recognized as a rare disease in many countries, including the USA, China, Japan, and South Korea.

Sepiapterin (Sephience) is a new molecular entity that is structurally identical to biologically produced sepiapterin [16, 17]. Sepiapterin exerts its pharmacological effects through dual mechanisms: (1) as an endogenous precursor to naturally occurring 5,6,7,8‐tetrahydrobiopterin (BH₄) in the pterin salvage pathway and (2) as a pharmacologically active chaperone that stabilizes the native conformation of PAH, which confers its resistance to thermal unfolding and results in enhanced activity and prolonged function [18, 19]. Clinical trials have demonstrated that following oral administration of sepiapterin, the major active metabolite present in systemic circulation is BH₄ [16, 20, 21, 22]. The maximum plasma concentration (C _max) and area under the curve from time 0 to time 24 h (AUC_0–24h) of sepiapterin were generally < 2% of those observed for BH₄ [16, 20, 22]. Additionally, sepiapterin concentrations were measurable only at a few time points, primarily around its time to maximum plasma concentration (T _max) at approximately 1 to 3 h post‐dose and declined to below the lower limit of quantitation (LLOQ, 0.5 ng/mL) generally by 12 h. BH₄ exhibits a terminal half‐life of approximately 5 h, with plasma concentrations generally reducing to near endogenous baseline levels by 24 h after oral sepiapterin administration [20].

BH₄ is an essential cofactor for several enzymes, including PAH, tyrosine hydroxylase (TH), tryptophan hydroxylase (TPH), alkylglycerol monooxygenase (AGMO), and nitric oxide synthase (NOS) [19]. Like sepiapterin, BH₄ also stabilizes the conformation of PAH and enhances its thermal stability and activity [23]. The efficacy and safety of sepiapterin for the treatment of patients with PKU have been demonstrated in the completed phase III pivotal global clinical study and the open‐label extension study [17, 24]. Recently, registration applications of sepiapterin were approved in Europe (June 2025), the USA (July 2025), and multiple additional countries worldwide for the treatment of HPA in patients with PKU [25, 26].

In the sepiapterin clinical development, no healthy volunteers or patients identified as Chinese or Korean participated in the clinical studies. However, the PK characteristics of sepiapterin were evaluated in 18 healthy Japanese volunteers and 15 Japanese patients with PKU [22]. In phase III studies, the majority of participants were White: 142 of 157 participants (90%) in the pivotal study and 142 of 169 participants (84%) in the ongoing open‐label extension study, as of the data cutoff date of September 2, 2024 [17, 24]. No Asian patients participated in the pivotal study, whereas 15 Asian patients (8.9%; all Japanese) participated in the extension study. In the absence of ethnic bridging studies, model‐based analysis of population PK and PK/PD (pharmacodynamic) is frequently used to assess ethnic sensitivity [27, 28]. However, this modeling approach is not applicable in the absence of PK and PD data for the targeted ethnic population, as it involves developing a model to describe the PK and/or PD characteristics of the drug using data from the overall clinical trial population, including the targeted ethnic population and identifying potential ethnic differences among ethnic populations [29].

This study presents an approach that combines an ethnic sensitivity assessment, conducted in accordance with the ICH E5 guidelines, with predictions of the geometric mean ratios (GMRs) for BH₄ exposure in Japanese, Chinese, and Korean populations relative to the White as a valuable tool to support the acceptance of foreign clinical data with no ethnic bridging studies for regulatory submissions.

2. Methods

2.1. Ethnic Sensitivity Assessment

The intrinsic and extrinsic factors that may contribute to ethnic differences were evaluated in accordance with ICH E5 guidelines. The potential factors sensitive to ethnic differences are outlined in Appendix D of the guideline [1].

2.2. Correlation of BH₄ C _max and AUC_{0 –24h} and Ethnic Group

Based on the results of the ethnic sensitivity analysis, breast cancer resistance protein (BCRP) c.421C>A variation was identified as the primary intrinsic factor affecting exposure to BH₄. To assess the potential impact on the PK of sepiapterin, the relative BH₄ exposures (C _max and AUC) ratios in participants carrying mutated BCRP c.421 genes (c.421CA or c.421AA) to the wild‐type carriers (BCRP c.421CC) were derived. The relative BH₄ exposure (C _max and AUC) ratio for subjects carrying the BCRP c.421CC gene was set as the baseline, with the value of 1. The relative BH₄ exposure ratios of BCRP c.421CA to c.421CC following oral dosing of sepiapterin were previously reported to be 1.36 and 1.39 for C _max and AUC_0‐last, respectively [22]. The relative BH₄ exposure ratios of BCRP c.421AA to c.421CC were not readily available from clinical data. However, they can be derived by linear extrapolation based on the BH₄ c.421CA to c.421CC ratios and the relative ratios of rosuvastatin, a commonly used clinical indicator for BCRP substrate, c.421AA to c.421CC and c.421CA to c.421CC, as described in Equation (1) below.

$$\begin{array}{c}\mathrm{BH}4\mathrm{c}421\mathrm{AA}/\mathrm{CC}\text{ratio}=\mathrm{BH}4\mathrm{c}421\mathrm{CA}/\mathrm{CC}\text{ratio}*\hfill \\ \left(\frac{\text{rosuvastatin}\mathrm{c}421\mathrm{AA}/\mathrm{CC}\text{ratio}}{\text{rosuvastatin}\mathrm{c}421\mathrm{CA}/\mathrm{CC}\text{ratio}}\right)\hfill \end{array}$$ (1)

The relative BH₄ exposure score for a specific population was derived by summing all products of the mutation frequency and the relative BH₄ exposure ratio of all individual genotypes for the specific mutation, as described in Equation (2) below.

$$\begin{array}{c}\text{Relative}\mathrm{BH}4\text{exposure score}\hfill \\ =\sum _{i=1}^n\text{frequency of}\mathrm{c}421\text{mutation}*\text{relative}\mathrm{BH}4\text{exposure ratio}\hfill \end{array}$$ (2)

where i and n represent the individual genotype and the total number of genotypes of a specific mutation.

The ratio of BH₄ exposure score for each ethnic group (i.e., Japanese, Chinese, and Korean) relative to White reflects ethnic differences, expressed as GMRs. The relative confidence interval (CI) of predicted BH₄ score ratios was estimated as per Equation (3) below.

$$\left[U,L\right]=\mathrm{Exp}\left[\mathrm{Ln}\left(\text{mean ratio}\right)\pm t*\mathrm{\sigma }/\sqrt{N}\right]$$ (3)

where U and L are the upper and lower bounds of the CI, t = 1.96 for a two‐sided 90% CI, σ ² is the intrasubject variance estimated based on clinical study results, and N is the sample size of the corresponding clinical study.

2.3. Validation of Correlation Predictions

Ethnic differences in the PK of orally administered sepiapterin between healthy Japanese and non‐Japanese individuals were assessed previously [22]. The observed PK differences between Japanese and non‐Japanese were compared with predictions based on the frequency of the BCRP c.421C>A variant in Japanese, using Equation (2) for validation. The prediction methodology would be considered reliable if the observed GMRs of BH₄ C _max and AUC_0–24h were contained within the 90% CI of predictions. Predicted results for Chinese and Korean populations would be deemed reliable if predictions for Japanese were successfully validated.

3. Results

3.1. Ethnic Sensitivity Assessment

3.1.1. Pharmacokinetics and Pharmacodynamics

The maximum recommended dosage of sepiapterin for the treatment of PKU patients is 60 mg/kg once daily. Following oral administration in humans, a saturable absorption of sepiapterin was observed [30]. The C _max and AUC_0–24h of the major systemic circulating metabolite BH₄ increased approximately dose‐proportionally at doses below 20 mg/kg, but far less than dose‐proportionally at doses above 20 mg/kg [20, 30]. Within the recommended sepiapterin dose range for PKU treatment, a flat PD curve for safety and efficacy was observed. No serious treatment‐related treatment‐emergent adverse events (TEAEs) were reported in PKU patients at sepiapterin doses up to 60 mg/kg [24] or in healthy volunteers at doses up to 120 mg/kg [30]. The incidence of TEAEs was comparable among PKU patients receiving placebo and those administered sepiapterin 20, 40, or 60 mg/kg once daily [17], as well as in healthy volunteers who received placebo or sepiapterin at 60 or 120 mg/kg [30]. During the placebo‐controlled part of the pivotal phase III study [17], PKU patients who were treated with sepiapterin at doses of 20, 40, and 60 mg/kg once daily exhibited substantial reductions in blood Phe levels compared to placebo, while the magnitude of the reduction in blood Phe levels increased moderately with higher doses [17]. Overall, the dose–response curves for both the safety and efficacy of sepiapterin appeared relatively flat.

Additionally, a population PK and PK/PD model was developed to assess the efficacy of sepiapterin, as measured by blood Phe reduction following the initiation of sepiapterin treatment [31]. The model incorporated PK data from both healthy volunteers and patients of all ages with PKU, as well as PD data from patients who participated in the phase III studies. In the final PK and PK/PD models, the impact of intrinsic and extrinsic factors was investigated to identify those with significant effects. In the PK model, it was found that Asian served as a covariate for the fraction of absorption, and weight was a covariate for both clearance and volume of distribution [31]. Once Asian was incorporated as a covariate into the final population PK model, further differentiation between Japanese and non‐Japanese participants was not significant. This indicates that Japanese individuals could not be distinguished from other Asian participants (whose specific ethnic backgrounds were not identified during clinical trials). Furthermore, the population PK/PD model demonstrated that neither ethnic nor race‐related factors significantly impacted the efficacy response, suggesting that sepiapterin effectiveness is not influenced by ethnicity [31].

3.1.2. Metabolism

Following oral administration, sepiapterin is primarily absorbed in the intestine [32]. Once absorbed, it is rapidly distributed to various tissues, including the central nervous system, and can move freely into and out of cells [33]. The metabolic pathway of sepiapterin has been extensively studied [19, 34, 35, 36, 37]. Exogenously administered sepiapterin is rapidly and extensively converted to BH₄ in vivo by a unidirectional, two‐step reaction via sepiapterin reductase (SPR) and dihydrofolate reductase (DHFR), following the salvage pathway of BH₄ biosynthesis (Figure 1). Carbonyl reductase is also involved in the first step of sepiapterin reduction to form the intermediate metabolite, 7,8‐dihydrobiopterin (BH₂) [19]. The conversion is rapid and extensive, as evidenced by the resultant sepiapterin exposure (measured according to C _max and AUC_0–24h) is typically < 2% of that of BH₄ [16, 20, 22]. The T _max of sepiapterin was approximately 1–3 h, while the T _max for BH₄ was approximately 4–5 h [16, 20]. The rapid and extensive conversion of sepiapterin to BH₄ across all sepiapterin doses (2.5–120 mg/kg) indicates that this process is neither rate‐ nor capacity‐limited. Hence, this process is unlikely to be influenced by ethnic differences, although there is a lack of studies on genetic polymorphisms in SPR and DHFR.

FIGURE 1.

Metabolic pathway of sepiapterin. Abbreviations: CAR, carbonyl reductase; DHFR, dihydrofolate reductase; SPR, sepiapterin reductase; XOR, xanthine oxidoreductase.

Similar to endogenous BH₄, BH₄ formed from exogenous sepiapterin is oxidized during catalytic aromatic amino acid hydroxylation while serving as a coenzyme for PAH, TH, TPH, AGMO, and NOS. Some metabolites formed in the process, such as BH₄‐4α‐carbinolamine, can be recycled to regenerate BH₄ via pterin‐4α‐carbinolamine dehydratase and dihydropteridine reductase [19]. Additional metabolites, such as BH₂, xanthopterin, isoxanthopterin, 7,8‐dihydroxanthopterin, lumazines, and other pteridine derivatives can be formed through enzymatic or non‐enzymatic pathways [34, 35, 36]. Therefore, the metabolic clearance of BH₄ is unlikely to be substantially altered by genetic polymorphism of a specific enzyme due to the co‐existence of multiple enzymatic and non‐enzymatic metabolic pathways.

In vitro studies have been conducted to assess transporters that may contribute to the absorption or uptake of sepiapterin and BH₄ in Madin‐Darby canine kidney II cells. Sepiapterin was identified as a substrate of BCRP, but not of P‐gp or BSEP [21]. BH₄ was identified as a substrate of both P‐gp and BCRP [21]. Within clinically relevant concentrations, it is unlikely that the transporters MATE1, MATE2‐K, OAT1, OATP1B1, OATP1B3, OCT1, and OCT2 are involved in the uptake of sepiapterin or BH₄.

Genetic polymorphisms in both P‐gp and BCRP have been reported [38]. Since the conversion of sepiapterin to BH₄ occurs exclusively intracellularly after sepiapterin uptake [33], the polymorphism of P‐gp is not expected to affect sepiapterin PK. There are over 80 naturally occurring single nucleotide polymorphisms of BCRP [38], the most widely studied being the c.421 variant, which is highly expressed in the gastrointestinal tract and has the most significant clinical impact [39, 40]. The c421C>A variant exhibits high racial differences, with the highest frequency in Asian populations, and is more common in Latin Americans with predominantly European and Native American ancestry compared to those with Afro‐Caribbean ancestry [38]. The c421C>A variant is associated with lower BCRP protein expression and, consequently, higher absorption due to the loss of efflux through BCRP, which results in higher oral drug exposures. Hence, it is expected that there will be racial differences in sepiapterin and BH₄ exposures.

3.1.3. Intersubject Variability of PK

In healthy volunteers, the intersubject variability (percentage of coefficient of variance [%CV]) for BH₄ C _max and AUC_0–24h (n = 31) was 25.5% and 27.3%, respectively, when sepiapterin 60 mg/kg was administered with a high‐fat, high‐calorie diet [30]. In patients with PKU aged 2 years or older, the intersubject variability for BH₄ C _max and AUC_0–24h (n = 35) was 44.5% and 60.2%, respectively, when sepiapterin 60 mg/kg was administered with a patient‐individualized Phe‐restricted diet [24]. The mean fat content in the Phe‐restricted diet is comparable to that of the standard low‐fat, low‐calorie diet [41].

3.1.4. Protein Binding and Food Effects

Both sepiapterin and BH₄ exhibit low protein binding. In vitro studies determined that sepiapterin was bound to plasma proteins at a mean ratio of 15.4% in the presence of 0.1% dithiothreitol in the concentration range of 0.1–10 μM [26]. The mean binding of BH₄ to plasma proteins ranged from 24.1% to 41.3% in the concentration range of 2–15 μM in the presence of 0.5% β‐mercaptoethanol [26].

Higher BH₄ C _max and AUC_0–24h are expected when sepiapterin is administered with food. In healthy volunteers, a low‐fat, low‐calorie diet increased BH₄ C _max values 69%–72% higher and AUC_0–24h values 62%–73% higher at sepiapterin doses of 20 or 60 mg/kg. When administered with a high‐fat, high‐calorie diet, BH₄ C _max values were 221%–226% higher, and AUC_0–24h values were 251%–284% higher at sepiapterin doses of 20 or 60 mg/kg [20].

During the treatment of patients with PKU, all individuals are instructed to adhere to their personalized Phe‐restricted diet [25, 26]. The composition of the Phe‐restricted diet exhibits a mean fat content comparable to that of a standard low‐fat, low‐calorie diet [41]. Hence, the dietary differences among various ethnic groups are considered to have minimal impact on BH₄ exposure. This conclusion was supported by the observation of a similar magnitude of increase in BH₄ exposure in both Japanese and White populations during clinical studies, when sepiapterin was administered with a low‐fat, low‐calorie diet compared with fasting conditions [20, 22].

3.2. Predicting BH₄ C _max and AUC_{0 –24h} Relative Ratios for Chinese, Korean, and Japanese Versus White Individuals

Based on the ethnic sensitivity analysis, the frequency of the c.421C>A genetic polymorphism of BCRP of each ethnic group is identified as the main factor resulting in differences in BH₄ exposures. The frequency of the c.421C>A variant is higher in Asian populations than in the White population (Table 1) [42]. The relative BH₄ C _max and AUC_0–24h scores for White, Japanese, Chinese, and Korean individuals were derived using Equation (2).

TABLE 1.

Frequency of BCRP c.421C>A variant in White, Japanese, Chinese, and Korean populations.

BCRP genotypes	White	Japanese	Chinese	Korean
CC	0.82	0.50	0.43	0.52
CA	0.17	0.41	0.45	0.40
AA	0.01	0.09	0.12	0.08

Note: Li and Barton 2018 [42].

Abbreviation: BCRP, breast cancer resistance protein.

The ratios of BH₄ C _max and AUC in those with BCRP c.421CA and AA variants relative to CC are summarized in Table 2. The ratios for BH₄ C _max and AUC in BCRP c.421CA genotype to CC of 1.36 and 1.39 were obtained from a clinical study reported earlier [21]. In the same study, the rosuvastatin C _max and AUC_0‐last in participants carrying the c.421CA gene were 1.60‐fold and 1.61‐fold of those carrying the c.421CC gene. The results for rosuvastatin were comparable to reported studies in Chinese healthy adults, which found that the rosuvastatin C _max and AUC_0–72h ratios in participants carrying the c.421C>A mutation (n = 5 for c.421CA and n = 2 for c.421AA) were 1.94‐fold and 1.76‐fold of those carrying the c.421CC gene (n = 7) [40]. The slightly higher increase of rosuvastatin in the report was attributed to the inclusion of participants with c.421AA, as it is known that exposures increase in the ascending order of c.421CC, c.421CA, and c.421AA for most BCRP substrates [43]. Hence, the rosuvastatin C _max and AUC_0‐last ratios in carriers with c.421CA to CC of 1.60‐fold and 1.61‐fold were used for the calculation. In a separate clinical study in healthy Chinese volunteers, rosuvastatin C _max and AUC_0–48h ratios in carriers of c.421AA to carriers of the c.421CC were 3.27 and 2.64, respectively [43]. The ratios of BH₄ C _max and AUC in carriers of c.421AA to carriers of c.421CC were estimated to be 1.65 and 2.28, based on the linear extrapolation of relative ratios of BH₄ and rosuvastatin, as specified in Equation (1).

TABLE 2.

Relative BH₄ and rosuvastatin C _max and AUC ratios in participants with various BCRP c.421 variants.

BCRP genotypes	BH4		Rosuvastatin
	C max	AUC	C max	AUC
CC a	1	1	1	1
CA b	1.36	1.39	1.60	1.61
AA c	1.65	2.28	1.94	2.64

Abbreviations: BCRP, breast cancer resistance protein; BH₄, 5,6,7,8‐tetrahydrobiopterin.

^a The wild‐type BCRP c.421CC was set to the value 1, as the baseline.

^b Results from adult healthy participants from a clinical study [22].

^c Ratios for rosuvastatin BCRP c.421AA versus c.421CC are from Wan, et al. 2015, [43] ratios for BH₄ were derived based on ratios of BCRP c.421CA to CC based on Equation (1).

The relative BH₄ exposure scores for each ethnic group were then calculated using Equation (2), with the relative ratios of BH₄ exposure in participants carrying BCRP c.421CC, CA, and AA genes (Table 2), and the frequency distribution of the mutation in White, Japanese, Chinese, and Korean populations (Table 1). The maximum intrasubject variance (σ ²) for BH₄ C _max and AUC_0–24h in healthy participants was estimated to be 0.0586 (intrasubject CV%, 24.6%) based on clinical data from the relative bioavailability study at sepiapterin doses of 20 and 60 mg/kg [20]. The 90% CIs of estimated GMRs between ethnic groups were calculated using Equation (3), where t = 1.96 for two‐sided 90% CI and the sample size N of 6 (per dose per group) as the clinical Japanese ethnic bridging study [22]. The value of $t*\mathrm{\sigma }/\sqrt{N}$ was estimated to be 0.1938, slightly less than the frequently used value of 0.223 for the common assumption of 25% intrasubject CV% (Table 3). The BH₄ C _max and AUC_0–24h GMRs (90% CI) for Japanese to White populations were predicted to be 1.13 (0.93, 1.37) and 1.18 (0.97, 1.43), respectively (Table 3). The corresponding GMRs for Chinese to White populations were 1.16 (0.96, 1.41) and 1.23 (1.02, 1.50), and for Korean to White populations were 1.12 (0.92, 1.36) and 1.17 (0.96, 1.42). BH₄ exposures for Japanese, Chinese, and Korean populations are predicted to be moderately higher compared to White populations. However, given the small magnitude of these ratios and the flat safety and efficacy response curves for sepiapterin, dose adjustment is not considered necessary for these Asian ethnic groups.

TABLE 3.

Relative BH₄ C _max and AUC_0–24h scores in participants from various ethnic groups and the relative ratios [mean (90% CI)] to White participants.

Parameter	White	Japanese	Chinese	Korean
Relative BH4 exposure scores
BH4 C max	1.068	1.206	1.240	1.196
BH4 AUC0–24h	1.079	1.275	1.329	1.258
Mean ratios to the White population (90% CI)
BH4 C max	1.0 (0.82, 1.21)	1.13 (0.93, 1.37)	1.16 (0.96, 1.41)	1.12 (0.92, 1.36)
BH4 AUC0–24h	1.0 (0.82, 1.21)	1.18 (0.97, 1.43)	1.23 (1.02, 1.50)	1.17 (0.96, 1.42)
Observed mean ratios of Japanese to non‐Japanese BH4 C max and AUC0–24h a
BH4 C max	NA		NA	NA
sepiapterin 20 mg/kg		1.14 (0.94, 1.37)
sepiapterin 40 mg/kg		1.29 (1.04, 1.59)
sepiapterin 60 mg/kg		1.29 (1.07, 1.56)
BH4 AUC0–24h	NA		NA	NA
sepiapterin 20 mg/kg		1.10 (0.92, 1.30)
sepiapterin 40 mg/kg		1.24 (1.02, 1.50)
sepiapterin 60 mg/kg		1.22 (1.03, 1.45)

Abbreviations: BCRP, breast cancer resistance protein; BH₄, 5,6,7,8‐tetrahydrobiopterin; CI, confidence interval; NA, not applicable.

^a Results in adult healthy participants from a clinical study [21].

3.3. Validation of Predictions

The reliability of predictions of BH₄ C _max and AUC_0–24h GMRs between Asian and White populations were validated by comparing the prediction with the clinical results from the Japanese ethnic bridging study [22]. In the clinical study, PK data were obtained from 18 Japanese and 18 non‐Japanese healthy adults, who were randomly assigned to receive a single dose of sepiapterin 20, 40, or 60 mg/kg (n = 6 per dose for each group) administered with a low‐fat low‐calorie meal [22]. The non‐Japanese participants comprised 15 White, 2 Black or African American, and 1 categorized as ‘Other’. The GMRs of natural logarithm‐transformed BH₄ exposures of Japanese to non‐Japanese at sepiapterin doses of 20, 40, and 60 mg/kg were 1.14, 1.29, and 1.29 for C _max and 1.10, 1.24, and 1.22, for AUC_0–24h, respectively. All GMRs were contained within the 90% CI of predictions (Table 3).

The consistency between predictions based on the BCRP c.421C>A variant frequency and clinical observations in the Japanese population confirmed the reliability of the predictions and supported the hypothesis that the BCRP c.421C>A variant is the main factor leading to ethnic differences in BH₄ exposures following sepiapterin oral administration. Hence, it was concluded that predictions based on the same methodology for Chinese and Korean populations are reliable.

4. Discussion

Understanding ethnic differences is crucial when extrapolating foreign clinical data to support NDAs in a different region. The ICH E5 guideline provides a general framework for assessing the acceptability of foreign clinical data for registration approval. Both intrinsic and extrinsic factors can lead to ethnic differences in the PK and PD of drugs. Genetic polymorphisms in drug‐metabolizing enzymes are a leading intrinsic factor influencing PK variability among ethnic groups, particularly in the CYP2D6, CYP2C, and CYP1A subfamilies. However, ethnic differences in CYP3A polymorphisms are inconsistent, likely due to significant intersubject variability [44, 45]. Additionally, differences in the frequency of genetic polymorphisms in transporters can have a significant clinical impact. Such differences are observed between Asian and European populations for OATP, P‐gp, and BCRP [38, 46]. Both P‐gp and BCRP are highly expressed in the gastrointestinal tract, and the related protein expression differences can lead to variations in drug absorption and distribution [38]. Extrinsic factors include differences in regional medical practices, tobacco and alcohol usage, exposure to sunlight and pollution, and compliance with prescribed medications [1]. Among all extrinsic factors, the effect of food is the most extensively studied, as food can affect both the rate and extent of oral absorption [45].

Despite various factors that may lead to differences in PK, the actual clinical observation of such differences is infrequent. A study of 114 eligible drugs approved between 2001 and 2008 in Japan found that only five drugs had C _max or AUC ratios outside the range of 0.5 to 2.0 [6]. A more recent study of 620 new molecular entities approved by the US Food and Drug Administration (FDA) between 2008 and 2023 indicated that only 5.0% reported ethnic/racial differences in PK [5]. For 25 drugs submitted for their first NDA in Japan between 2010 and 2018, only one was found to have ethnic differences in PK [47]. For a total of 22 NDAs in China between 2010 and 2018, phase I ethnic bridging studies were conducted for 10 compounds, and no ethnic differences in safety or PK were found in these 10 compounds [47]. This low frequency of ethnic differences in PK did not translate into a correspondingly low frequency of differences in approved doses between regions. A study of 137 NDAs in both Japan and the USA between 2001 and 2007 indicated that for 43 of them the dose was higher in the USA than Japan, and 40 out of these 43 had a dose ratio (USA to Japan) of ≥ 2.0 [48]. Multiple studies have found that there was no correlation between differences in the recommended dose and potential intrinsic factors or PK exposures (e.g., AUC) [5, 49]. The frequency of dose differences among ethnic groups varies by drug class; it is rare for anticancer and antiviral drugs but common for cardiovascular drugs [48]. These findings suggest that other factors, such as regional medical practices, safety, and tolerability, may influence approved doses more than PK parameters [5]. Hence, in addition to examining ethnic differences in PK and PD, it is important to assess the overall safety and efficacy of a drug, as well as the gradient of safety and efficacy responses to exposure. These factors are essential for the acceptability of foreign clinical data and the determination of the approved dose. The growing focus on this comprehensive approach is reflected in the 2020 guideline issued by the China NMPA [8]. According to this guideline, drugs already approved outside of China can be considered for approval even in the absence of clinical data from Chinese populations for rare diseases, provided no significant ethnic sensitivity is identified [8, 50].

In this study, it has been demonstrated that the genetic polymorphism of the BCRP c.421C>A variant is the main intrinsic factor leading to ethnic differences in the PK of sepiapterin for the treatment of patients with PKU. Differences in BH₄ exposures among Chinese, Korean, and Japanese populations compared with the White population can be predicted based on the frequency of the BCRP c.421C>A mutation, and such predictions have been validated by clinical data from the phase I Japanese and non‐Japanese ethnic bridging study [22]. No substantial difference in BH₄ exposures is predicted for Chinese, Korean, and Japanese versus Whites, and the efficacy and safety responses of sepiapterin are relatively flat in the recommended therapeutic dose range. Overall, sepiapterin safety and efficacy are insensitive to ethnic differences.

The global prevalence of PKU is approximately 1 in every 23,930 births, and 1:125,000 in Japan; 1:25,000 in the USA, 1:15,924 in China, and 1:41,000 in Korea [15]. PKU is recognized as a rare disease in the EU, the USA, Japan, China, South Korea, and several other countries, and sepiapterin is designated as an orphan drug for treatment of PKU patients in the EU, the USA, Japan, South Korea, and several other countries. There were two existing treatments for PKU prior to sepiapterin approval, oral administration of sapropterin dihydrochloride (BH₄ supplements) and enzyme substitution by daily injection of pegvaliase, a PEGylation of the Anabaena variabilis phenylalanine ammonia lyase enzyme. However, only approximately 20% of PKU patients respond to sapropterin treatment [11, 12], while pegvaliase is restricted to patients aged 15 years or above and requires a long‐term titration to identify the right dose. Pegvaliase is also associated with a high rate of adverse events including anaphylaxis. This highlights the urgent unmet medical need for improved treatments for PKU patients. Sepiapterin has demonstrated superior efficacy to sapropterin and is efficacious in a broader range of PKU patients [12, 15, 17, 51]. Sepiapterin is approved for the treatment of PKU patients of all ages (EU and Australia) or 1 month or older (USA, Japan, and many other countries recently). Expanding access to sepiapterin for additional racial and ethnic populations is critical to addressing this unmet need. Waiving clinical ethnic bridging studies for these populations can significantly reduce time delays enabling faster access to treatment. This approach is supported by the long‐term safety and efficacy data demonstrated in PKU patients, including Japanese patients, from the ongoing clinical trials [24] and the findings from this study.

5. Conclusions

The ethnic sensitivity analysis of sepiapterin conducted in this study identified the BCRP c.421C>A genetic polymorphism as the primary intrinsic factor leading to PK differences among ethnic groups. A relationship was established between PK differences in Chinese, Korean, and Japanese populations compared to the White population and the frequency of BCRP c.421C>A mutations. This relationship was validated using clinical data from a phase I ethnic bridging study involving both Japanese and non‐Japanese individuals [22]. Predictions based on this model indicated that the BH₄ C _max and AUC_0–24h in Chinese and Korean populations were slightly higher than those in the White population yet remained comparable to the levels observed in the Japanese population. Given the relatively flat exposure‐response relationship of sepiapterin with safety and efficacy, this mild increase in BH₄ exposure in Chinese and Korean populations was deemed clinically irrelevant. Consequently, the same sepiapterin dose regimen was proposed for Chinese, Korean, and White populations. The results from the ethnic sensitivity analysis, the validated predictions for relative BH₄ exposures in Chinese and Korean based on BCRP c.421C>A mutation frequency, the flat dose response relationship of sepiapterin safety and efficacy, the rarity of PKU disease and highly unmet medical need, collectively support waiving the requirement to conduct an ethnic bridging study in these populations for regulatory filing.

Author Contributions

L.G. wrote the manuscript. L.G., N.S., and R.K. designed the study. L.G. performed the research. L.G. analyzed the data.

Funding

This research was supported by funding from PTC Therapeutics Inc.

Ethics Statement

The authors have nothing to report.

Consent

The authors have nothing to report.

Conflicts of Interest

All authors are employees of PTC Therapeutics Inc. and may hold shares in PTC Therapeutics.

Data Availability Statement

All source data used in this study are available publicly.