Safety, Tolerability, Pharmacokinetics, Food Effect of Ribitol, and Its Effect on QTcF in Healthy Adults: First‐in‐Human, Randomized, Double‐Blind (Sponsor Unblinded), Placebo‐Controlled Studies

Abstract

Limb‐girdle muscle dystrophy Type R9 (LGMDR9), also known as LGMD Type 2I, is a rare genetic disease caused by partial loss of function of fukutin‐related protein (FKRP) enzyme which glycosylates alpha‐dystroglycan, thereby stabilizing myocytes during contraction. Hypoglycosylation leads to progressive muscle injury and impaired function including loss of ambulation. Ribitol is an endogenous pentose alcohol and precursor to CDP‐ribitol, the substrate of FKRP. This first‐in‐human study demonstrated that ribitol was well tolerated when administered as single or multiple oral doses over 6 days to healthy adults. PK demonstrated dose‐proportional increases in exposure from 0.5 to 15 g (therapeutic dose 9 and 12 g BID for patients weighing >30 to ≤50 kg and >50 kg, respectively); t_½ was 9‐13 h. A high‐fat meal did not affect overall oral bioavailability; indicating ribitol may be taken without regard to food intake. A dedicated QT study using ribitol 21 g revealed no concentration‐dependent QTcF prolongation and clinically significant QTcF prolongation was excluded over the entire range of exposures in the study, up to 351.9 µg/mL. Assay sensitivity was demonstrated with the expected effect of moxifloxacin. These results support further development of ribitol for the treatment of LGMDR9.

Limb‐girdle muscular dystrophy Type R9 (LGMDR9), also known as LGMD Type 2I, is an ultra‐rare monogenic autosomal recessive neuromuscular disease caused by partial loss of function mutations in the fukutin‐related protein (FKRP) gene and marked by progressive debilitating and life‐threatening complications. ¹ , ² This condition results in deficient activity of the FKRP enzyme which functions as a ribitol‐5‐phosphate transferase and glycosylates alpha‐dystroglycan (αDG). ³ , ⁴ When adequately glycosylated, αDG is essential for anchoring muscle cells to the extracellular matrix thus protecting it from injury during use. ¹ Patients living with LGMDR9 display a marked reduction in αDG glycosylation compared to healthy controls. ¹ , ⁵ This results in a lack of muscle stabilization during contraction and continuous muscle injury that overwhelms repair mechanisms and leads to progressive muscle loss and weakness. Disease onset and progression vary by the type of mutation, but most patients lose ambulation during adulthood, and some develop cardiomyopathy and/or diaphragmatic insufficiency requiring assisted ventilation. ⁶

There are currently no approved therapies that modify the course of LGMDR9, and interventions are limited to supportive care including physical therapy to maintain mobility and prevent contractures, orthopedic interventions for scoliosis or contractures, and nutritional and general rehabilitation. Respiratory support and drug therapy for cardiac systolic dysfunction are added as needed.

Ribitol is an endogenous pentose alcohol which after conversion to cytidine‐5‐diphosphate‐ribitol (CDP‐ribitol) becomes the substrate for the enzymes fukutin (FKTN) and FKRP. These enzymes utilize CDP‐ribitol to add ribitol‐5‐phosphate to αDG, a crucial step for its proper glycosylation, enabling the protein to stabilize myocytes during contraction. The exogenous administration of ribitol to patients with LGMDR9 provides an opportunity to maximize the residual FKRP activity present in all patients and increase αDG glycosylation, thus addressing the disease pathophysiology at its root. Accordingly, ribitol is being developed by BridgeBio Pharma Inc. and ML Bio Solutions Inc. as an oral therapy to address the progression of muscle loss. Preclinical studies documented the favorable safety and tolerability of ribitol at doses shown histologically to increase αDG glycosylation and improve muscle performance in a mouse model of LGMDR9. ⁷

Here we report the results of the first‐in‐human study designed to evaluate the safety, tolerability, and pharmacokinetics (PK) of ribitol when administered orally as either single ascending doses (SAD) or multiple ascending doses (MAD) to healthy adults; the food effect (FE) on the PK of a single oral dose of ribitol; and the effect of ribitol on the QTc interval corrected using the Fridericia method (QTcF) assessed in a dedicated QT study, the concentration‐QTcF analysis being the primary endpoint.

Methods

Participants

All studies were conducted in accordance with the ethical principles described in the Declaration of Helsinki and consistent with the International Conference on Harmonization for Good Clinical Practices guidelines and applicable local and regulatory requirements. Advarra, Inc. Institutional Review Board approved the research protocols. Written informed consent was obtained before any study‐related procedures were conducted.

The studies enrolled healthy men and women ages 18 to 65 years (19 to 65 in Nebraska) and a BMI of 18 to 32. They were screened to ensure they were in good health as determined by medical and medication history, physical examination, hematology and chemistry laboratory tests, HIV and hepatitis screening, and electrocardiograms (ECGs). Participants were not eligible if they were pregnant or breast feeding, had present or past significant medical illness, or were using any prescription medicine other than contraceptives or occasional acetaminophen within 4 weeks of dosing and any investigational medication within 12 weeks of dosing.

Study Design

This first‐in‐human randomized, double‐blind (sponsor unblinded), placebo‐controlled, single center SAD/MAD study was conducted between June 2020 and April 2021 at Celerion in Lincoln, NE; the FE study was conducted between August 2024 and November 2024 at Celerion in Tempe, AZ; the dedicated QT study was conducted between August 2024 and November 2024 at Celerion in Tempe, AZ.

In the FIH and FE studies, ribitol was administered as a 200 mg/mL solution in purified water (United States Pharmacopeia [USP]); placebo consisted of a 0.35 mg/mL sucralose oral solution in purified water (USP). Both ribitol and placebo were compounded at the Celerion clinical pharmacy; both were similar in all aspects including volume and color. In the QT study, ribitol was administered as seven sachets containing 3 g ribitol granules or its placebo dissolved in purified water (USP). Moxifloxacin was supplied as 400 mg encapsulated moxifloxacin hydrochloride tablets. Moxifloxacin matching placebo capsules contain microcrystalline cellulose (National Formulary [NF]). Immediately prior to dosing, participants were given one strip of Listerine Cool Mint Pocketpaks as a taste‐masking agent before dosing to mask any flavor differences between ribitol and placebo.

For each SAD and MAD dose level, a cohort of eight participants was randomized to receive ribitol or its placebo in a 3:1 ratio. In the FE study, all 16 enrolled participants received ribitol. In the dedicated QT study, participants each received ribitol, placebo, and moxifloxacin in a crossover fashion.

First‐in‐Human, Single Ascending Dose Part

Safety, tolerability, and PK of single doses of an oral ribitol solution or its placebo administered after a 10 h overnight fast and followed by an additional 4 h fast were assessed. The starting dose was 0.5 g orally; subsequent SAD dose levels were determined by a safety review committee who reviewed all safety data through at least 72 h after the last dose and PK data through at least 24 h after the last dose. The following doses were studied: 0.5, 1.5, 3, 6, 9, 12, and 15 g. Each dosing group was preceded by a sentinel group of two participants, who were dosed 24 h prior to the rest of the cohort, one each receiving ribitol and placebo. The SAD portion included a preliminary FE portion where one cohort received 3 g ribitol with and without food based on whose results the PM dose of the BID administration in MAD was given without regard to food (see later).

First‐in‐Human, Multiple Ascending Dose Part

Safety, tolerability, and PK of a 6‐day course of an oral ribitol solution or its placebo were assessed at doses of 1.5 g once daily (QD), 3 g QD, 3 g every 12 h (Q12H), 6 g Q12H, and 9 g Q12H. The last dose for all cohorts was in the AM of Day 6. For Q12H administrations, the AM doses on Days 1 and 6 were administered after a 10 h overnight fast and followed by an additional 4 h fast; all other doses were administered without regard to food. The starting dose and all subsequent dose levels were determined by a safety review committee who reviewed all safety data through at least 72 h after the last dose and PK data through at least 24 h after at least the first 3 SAD cohorts for the starting dose and after the previous MAD cohort for subsequent cohorts.

Food Effect

Safety, tolerability, and PK of a single oral dose of ribitol administered after a 10 h overnight fast and after a standardized high‐fat breakfast were assessed in 16 participants using an open label, two‐period, randomized sequence design. ⁸ There was a 96‐h washout period between the two doses. In both dosing periods a 4‐h fasting period was observed after dosing. The meal consisted of two fried eggs, two strips of bacon, two slices of toast with butter, two hash brown patties, and 8 ounces of whole milk to provide approximately 800‐1000 kcal, about 50% of calories from fat, 15% from protein, and 35% from carbohydrates.

QT Assessment

This was a two‐part study. Part 1 was an open label administration of a single 21 g ribitol dose to assess the safety, tolerability, and PK of this previously unstudied, supratherapeutic dose in six healthy adults. Part 2 consisted of a double‐blind (participants blinded), randomized, placebo‐ and positive‐controlled 3‐way crossover study in 28 healthy adults. Participants each received a single dose of (1) ribitol placebo plus a moxifloxacin placebo, (2) ribitol 21 g plus a moxifloxacin placebo, and (3) moxifloxacin 400 mg plus a ribitol placebo. The sequence was randomized and drug administrations were separated by a 4‐day washout. Participants remained in the clinical studies unit throughout the study. Continuous 12‐lead ECG recordings were obtained using M12R continuous 12‐lead digital recorders and the M12A Enterprise Holter System Client (Global Instrumentation, LLC, Manlius, NY). Participants were kept resting in a supine position for at least 10 min prior to each PK time point coinciding with ECG extractions.

Safety Assessments

Safety assessments were based on treatment‐emergent adverse events (TEAEs) reported throughout the study, vital sign measurements, physical examinations, clinical laboratory tests, and ECGs assessed at predefined time points. TEAEs were coded using the Medical Dictionary for Regulatory Activities Version 23.0 (SAD/MAD) and 24.1 (FE). Additional safety assessments in MAD included the Columbia Suicide Severity Rating Scale (C‐SSRS) scale which was assessed at baseline and at the end of study visit on Day 10.

Pharmacokinetics

Venous blood samples were collected on Day 1 predose and at 0.25, 0.5, 1, 1.5, 2, 3, 4, 6, 8, 12, 18, 24, 36, 48, and 144 h after dosing for SAD and FE; for MAD, collection time points were Day 1 predose and at 0.25, 0.5, 1, 1.5, 2, 3, 4, 6, 8, 12, and 18 h after the first dose. On Days 2 to 5, predose and at 0.5 after dosing; on Day 6, predose and at 0.25, 0.5, 1, 1.5, 2, 3, 4, 6, 8, 12, 18, 24, 36, 48, and 96 h after the last dose. In the QT study, collection time points for ribitol and moxifloxacin plasma concentrations were predose and 0.5, 1, 1.5, 2, 3, 4, 6, 8, 12, 24, 36, and 48 h after dosing. A 24‐h urine collection was obtained at steady state starting with the administration of the last dose in MAD.

Ribitol extracted from plasma was quantified by gas chromatography‐mass spectrometry (GC‐MS/MS) using a standard curve prepared in a surrogate matrix (6% SigMatrix Serum Diluent) and quality controls prepared in both plasma and the surrogate matrix. The dynamic range of this assay is 20.0 to 10,000 ng/mL; concentrations higher than the upper limit of quantitation may have been accommodated by dilution with blank matrix. Acetonitrile precipitation and GC‐MS/MS was used to determine the concentration of the analyte present in the matrix. An aliquot of the extract was injected onto a GC‐MS/MS triple quadrupole mass spectrometer. The peak area of the product ion of the analyte was measured against the peak area of the product ion of the stable label internal standard (¹³C₅‐ribitol). A calibration curve spanning the curve range and containing at least six concentrations in duplicate was used to quantify the analyte concentration.

Plasma concentrations of moxifloxacin were determined using a validated LC/MS method. The analytical range (LLOQ–ULOQ) for moxifloxacin was 25.0‐25,000 ng/mL.

Urine samples were analyzed for ribitol using a validated LC‐MS/MS method. The method was validated for a range of 0.100 to 50.0 µg/mL based on the analysis of 50.0 µL of urine. Urine containing ribitol and the internal standard, ¹³C₅‐ribitol, was extracted using protein precipitation and analyzed by a Sciex API 4000 LC‐MS‐MS equipped with an HPLC column. The peak area of the m/z 151.2 → 71.1 ribitol product ion was measured against the peak area of the m/z 156.2 → 74.1 ¹³C₅‐ribitol internal standard product ion. Quantitation was performed using a weighted 1/x2 linear least squares regression analysis generated from calibration standards prepared on the day of extraction.

The PK analysis for plasma ribitol and moxifloxacin were performed using a noncompartmental method with validated software (Phoenix WinNonlin version 8.1 or 8.4 [Certara L.P., Princeton, NJ]. Baseline (endogenous) ribitol plasma concentrations were subtracted from all reported ribitol plasma concentration data prior to PK analysis. Estimated PK parameters included the maximum observed plasma concentration (C_max), time to reach C_max (T_max), concentration observed at the end of the dosing interval (C_trough), area under the plasma concentration–time curve (AUC) from time 0 to time of last measurable concentration (AUC_0‐last), AUC from time 0 to infinity (AUC_∞), AUC during the 12‐ or 24‐h dosing interval (AUC_0‐τ), elimination half‐life (t_½), the observed accumulation ratio based on C_max and AUC_0‐τ (RC_max and RAUC, respectively), the amount of ribitol excreted into the urine as a percent of dose and the renal clearance of drug.

Statistical Analysis

As is customary for first‐in‐human studies, formal statistical sample size considerations were not performed. The number of participants in each cohort is consistent with typical sample sizes in similar studies and was sufficient to allow clinical judgment of safety data to be used for dose escalation decisions and provide adequate assessment of the PK profile.

Ribitol and moxifloxacin plasma concentration data, derived PK parameters, and safety data were summarized using descriptive statistics. The PK population included all enrolled participants for whom at least one PK parameter of interest could be calculated. The safety population included all participants who received at least one dose of ribitol or placebo.

For the FE, a comparison of natural‐log (ln)‐transformed baseline‐adjusted PK parameters AUC_∞, AUC_0–last, AUC_0‐48, and C_max was made to evaluate the effect of the high‐fat meal. For each ln transformed PK parameter, the comparison of 12 g (fed) versus 12 g (fasted) was evaluated by performing an analysis of variance (ANOVA) model using PROC MIXED of SAS. The ANOVA model included treatment, sequence, and period as fixed effects and subject nested within sequence as a random effect. The geometric mean ratios (GMRs) and 90% confidence intervals (CIs) were expressed as a percentage relative to 12 g (fasted). The inferential results (least‐squares means [LSMs], difference between LSMs, and 90% CIs of the difference) were exponentiated to the original scale. Geometric LSMs, GMRs, 90% CIs of the GMRs, and intra‐subject CV% are presented.

Cardiodynamic Analyses

Up to 10 non‐overlapping 14‐s digital 12‐lead ECG tracings were extracted from the continuous 48‐h recordings at each PK time point. Baseline consisted of predose extractions obtained at 60, 45, and 30 min prior to dosing. ECG intervals were measured by eResearch Technologies, Inc., a Clario company, Philadelphia, PA in a blinded manner (blinded to treatment, visit, day, time point, and participant information) using their Early Precision QT (EPQT) technique.

All cardiodynamic statistical analyses were performed using the statistical software SAS for Windows Version 9.4 or higher (SAS Institute, Inc., Cary, NC). Missing ECG values were represented by an empty cell and no imputation was made. For all descriptive statistics for the absolute values and change‐from‐baseline values of continuous ECG parameters (i.e., heart rate, PR, QRS, and QTcF), data were summarized. For all modeling results of the by‐time‐point analysis of change‐from‐baseline values of continuous ECG parameters, number of participants (n), least squares (LS) mean, standard error (SE), and two‐sided 90% CI were included. Modeling results of the by‐time‐point analysis of placebo‐corrected change‐from‐baseline values also included LS mean, SE, and 90% CI.

The Fridericia formula was used to correct QT for heart rate as follows:

$$\mathrm{QTcF}\left(\mathrm{ms}\right)=\frac{\mathrm{QT}\left(\mathrm{ms}\right)}{\sqrt[3]{\frac{\mathrm{RR}\left(\mathrm{ms}\right)}{1000}}}$$

Change‐from‐baseline (Δ) for ECG parameters (including QTcF, heart rate, and QRS) was the difference between the post‐dose value and its corresponding baseline value. Placebo‐corrected change‐from‐baseline (ΔΔ) for ECG parameters was the difference in change‐from‐baseline ECG values between active treatment (or moxifloxacin) and placebo. For both the concentration‐QTc analysis, ΔQTcF was used as the dependent variable for calculation of model‐derived ΔΔQTcF.

The relationship between ΔQTcF and plasma concentrations of ribitol, the primary endpoints, was quantified using a linear mixed‐effects modeling approach. The model used ΔQTcF as the dependent variable, ribitol plasma concentrations as the explanatory variable (0 for placebo), centered baseline QTcF (i.e., baseline QTcF for an individual participant minus the population mean baseline QTcF for all participants within the same treatment period) as an additional covariate, study treatment (active  =  1 or placebo  =  0) and time (i.e., post‐dose time point on Day 1) as fixed effects, and random intercept and slope per participant. ⁹

The geometric mean of the individual plasma C_max values for participants in the ribitol dose group was determined. The model‐predicted effect population point estimate and its two‐sided 90% CI for placebo‐corrected ΔQTcF (ΔΔQTcF) (i.e., slope estimate  ×  concentration  +  treatment effect‐specific intercept) at this geometric mean C_max were obtained. If the upper bound of the two‐sided 90% CI (equivalent to the upper bound of the one‐sided 95% CI) of the model‐predicted QTc effect (ΔΔQTcF) was below 10 ms at the geometric mean ribitol C_max as well as clinically relevant plasma levels, it could be concluded that ribitol did not cause clinically significant QTc prolongation within the observed plasma concentration range. ¹⁰

Results

Participant Disposition and Demographics

A total of 54 participants were enrolled in SAD; the 9 and 15 g cohorts only enrolled seven participants due to the COVID epidemic. One placebo participant withdrew because of the number of blood draws and 53 participants completed the study. A total of 39 participants were enrolled in MAD and all completed the study; the 3 g Q12H cohort only enrolled seven participants due to the COVID epidemic. Sixteen participants were enrolled in the FE study; one participant discontinued early for personal reasons and only received ribitol in the fed state. For Part 1 of the dedicated QT study designed to evaluate the PK of ribitol 21 g, six participants were planned, and all completed the study per protocol. For Part 2, the protocol specified 27 participants; however, one participant did not complete placebo dosing and was replaced, for a total of 28 participants randomized and 27 completing the study.

The demographic characteristics were similar across groups. The sex distribution was well balanced with 59% males for SAD, 67% males for MAD, 50% male for FE, and 46% males for QT. Most participants were White, 57% for SAD, 59% for MAD, 94% for FE, and 86% for QT. African Americans accounted for 30% of participants in SAD, 36% in MAD, 6% in FE, and 11% in QT. Mean (range) age was 36.6 (20‐64) years for SAD, 41.3 (21‐65) years for MAD, 37.9 (22‐63) years for FE, and 34.9 (22‐52) in QT; mean BMI (range) was 25.6 (19‐32) for SAD, 26.2 (20‐32) for MAD, 27.2 (18‐32) for FE, and 26.0 (20‐32) for QT.

Safety

There were no serious adverse events (SAEs) or discontinuations due to an adverse event. In SAD, 40 TEAEs were reported by 20 (37%) participants. The incidence ranged from 33% at 6 g to 80% at 15 g; the incidence on placebo was 21%. The most common TEAE was headache (9% on ribitol and 7% on placebo); other common TEAEs included gastrointestinal AEs like abdominal discomfort/distention/pain/tenderness, constipation, diarrhea, and nausea (15% on ribitol and 5% on placebo). In MAD, 42 TEAEs were reported by 17 (44%) participants, ranging from 17% (3 g Q12H) to 67% (placebo Q12H and 6 g Q12H). The most common AE was vessel puncture site pain, reported by 5 (13%) participants overall. Gastrointestinal AEs (abdominal discomfort, diarrhea, dry mouth, dyspepsia, and flatulence) occurred in 67% of participants receiving 6 g Q12H and none of the other groups. Most TEAEs were mild in severity; 4 (3 on 1.5 g QD and 1 on placebo) were of moderate severity. In the FE study, 4 mild TEAEs were reported by 4 (25%) of the 16 participants; no single AE was reported by more than 1 participant. In Part 1 of the QT study, three participants reported a mild TEAE (two diarrhea and one headache). In Part 2, 33 TEAEs were reported by a total of 7 (25%) participants with 5 (19%) participants after ribitol, 4 (14%) participants after placebo, and 4 (15%) participants after moxifloxacin. The most common event was diarrhea (4 [14%] participants). Except for two events of diarrhea considered moderate, all AEs were considered mild. Lastly, no clinically significant changes were observed in laboratory parameters, vital signs, physical examinations, or ECGs were noted in SAD, MAD, FE, or QT studies.

Pharmacokinetics

Single‐Dose Pharmacokinetics (SAD)

Following a single dose of ribitol oral solution under fasted conditions in healthy adults, ribitol was rapidly absorbed as evidenced by a T_max of 1.5 h or less; thereafter, plasma concentration declined in a multiexponential manner (see Figure 1). Exposure as measured by C_max and AUC increased over the 0.5 to 15 g dose range in a mostly dose‐proportional fashion (Table 1) and interindividual variability was low for an oral drug. Mean t_½ ranged from approximately 9 to 13 h and did not appear to be dose dependent. Plasma ribitol concentrations in participants receiving placebo remained at or near baseline levels throughout the post‐dose sampling duration ranging from 42.5 to 56.5 ng/mL and thus being markedly lower than with active treatment.

Figure 1.

Mean plasma ribitol concentration–time profile (A) linear scale and (B) semi‐logarithmic scale. Arithmetic mean plasma ribitol concentrations are plotted over time and depict the exposures observed after single doses of a ribitol oral solution ranging from 0.5 to 15 g or placebo were administered to healthy adults in the fasted state. The concentrations are not adjusted for endogenous plasma ribitol levels, whose magnitude and stability over time are apparent in the semilogarithmic plot of the placebo group.

Table 1.

Summary of Baseline‐Adjusted Plasma Ribitol Pharmacokinetic Parameters Following Single Oral Doses of Ribitol Solution to Healthy Adults Under Fasted Conditions

Pharmacokinetic Parameters	0.5 g	1.5 g	3 g	6 g	9 g	12 g	15 g
Cmax (µg/mL)	7.520 (19.7) [n = 6]	28.66 (9.7) [n = 6]	44.64 (35.1) [n = 6]	80.36 (19.0) [n = 6]	117.6 (18.9) [n = 5]	120.1 (38.2) [n = 6]	221.0 (12.6) [n = 5]
Tmax (h)	0.515 (0.50, 1.00) [n = 6]	0.798 (0.58, 1.02) [n = 6]	0.854 (0.50, 1.01) [n = 6]	0.751 (0.50, 1.01) [n = 6]	0.530 (0.52, 1.51) [n = 5]	1.002 (1.00, 1.50) [n = 6]	0.983 (0.52, 1.50) [n = 5]
AUC∞ (µg h/mL)	18.97 (27.8) [n = 4]	62.96 (12.5) [n = 6]	108.8 (25.4) [n = 6]	207.0 (10.9) [n = 6]	344.8 (11.3) [n = 5]	346.7 (22.1) [n = 6]	542.9 (12.3) [n = 5]
AUC0–last (µg h/mL)	19.20 (24.1) [n = 6]	62.75 (12.6) [n = 6]	108.3 (25.4) [n = 6]	206.9 (10.9) [n = 6]	344.4 (11.4) [n = 5]	346.1 (22.1) [n = 6]	542.8 (12.3) [n = 5]
t1/2 (h)	11.180 ± 4.1456 [n = 4]	12.011 ± 2.7224 [n = 6]	12.842 ± 1.2183 [n = 6]	10.168 ± 1.6454 [n = 6]	8.649 ± 2.5455 [n = 5]	8.526 ± 1.1321 [n = 6]	9.332 ± 1.2085 [n = 5]

AUC, area under the plasma concentration–time curve; AUC_∞, AUC from time 0 to infinity with extrapolation of the terminal phase; AUC_0–last, AUC from time 0 to the time corresponding to the last quantifiable concentration; C_max, maximum observed plasma concentration; SD, standard deviation; t_1/2, terminal half‐life; T_max, time correspondent to the maximum observed plasma concentration.

Notes: AUCs and C_max values are presented as geometric mean (geometric CV%).

T_max is presented as median (min, max).

Other parameters are presented as arithmetic mean ± SD.

Multiple‐Dose Pharmacokinetics (MAD)

The PK parameters after the first dose (Day 1) and at steady state (Day 6) are summarized in Table 2. PK parameters after the first dose on Day 1 were largely consistent with those obtained after single doses. Visual inspection of daily trough concentrations indicated that steady state was achieved by approximately Day 3 in all cohorts. Steady‐state PK data on Day 6 revealed no accumulation for C_max and AUC_0‐τ with either QD or Q12H administration. Exposure increased in a dose‐proportional manner and t_½ averaged between approximately 11 and 17 h. Urinary ribitol excretion at steady state averaged between approximately 18% and 21% of the administered dose and was not dose dependent; renal clearance averaged between 4.5 and 5.6 L/h and did not appear dose dependent. Investigations into potential human metabolites of ribitol across several studies indicated that metabolism is not an important clearance mechanism. Ribulose, comprising only 3% of total parent, was the most abundant metabolite detected in human plasma.

Table 2.

Summary of Baseline‐Adjusted Plasma Ribitol Pharmacokinetic Parameters Following Administration of the First (Day 1) and Last (Day 6) Oral Dose of Ribitol

Pharmacokinetic Parameters	1.5 g QD	3 g QD	3 g Q12H	6 g Q12H	9 g Q12H
DAY 1
AUC(0‐τ) (µg h/mL)	61.83 (33.1) [n = 6]	87.15 (30.4) [n = 6]	99.35 (20.3) [n = 5]	254.8 (50.0) [n = 6]	344.8 (40.3) [n = 6]
Cmax (µg/mL)	30.86 (39.1) [n = 6]	33.76 (27.8) [n = 6]	41.52 (21.3) [n = 5]	87.26 (41.6) [n = 6]	135.9 (32.6) [n = 6]
Tmax (h)	0.795 (0.51, 1.00) [n = 6]	0.511 (0.50, 1.10) [n = 6]	0.999 (0.51, 1.00) [n = 5]	1.008 (0.50, 2.01) [n = 6]	0.551 (0.50, 1.01) [n = 6]
DAY 6
AUC(0‐τ) (µg h/mL)	59.78 (27.4) [n = 6]	94.97 (24.8) [n = 5]	109.4 (20.6) [n = 5]	250.0 (24.3) [n = 6]	336.6 (29.0) [n = 6]
Cmax, ss (µg/mL)	24.97 (22.8) [n = 6]	35.55 (20.1) [n = 5]	37.97 (23.4) [n = 5]	107.2 (33.6) [n = 6]	143.1 (32.1) [n = 6]
Tmax, ss (h)	0.5040 (0.500, 1.69) [n = 6]	0.5030 (0.500, 0.999) [n = 5]	0.5010 (0.500, 1.00) [n = 5]	0.7505 (0.257, 1.01) [n = 6]	0.5010 (0.500, 0.512) [n = 6]
t1/2 (h)	15.297 ± 4.0920 [n = 6]	16.710 ± 3.6225 [n = 5]	13.600 ± 3.6858 [n = 5]	12.539 ± 1.0306 [n = 6]	11.180 ± 2.1577 [n = 6]
RAUC	0.9774 ± 0.15702 [n = 6]	1.179 ± 0.074766 [n = 5]	1.105 ± 0.095967 [n = 5]	1.005 ± 0.22214 [n = 6]	0.9916 ± 0.18036 [n = 6]
RCmax	0.8251 ± 0.16639 [n = 6]	1.126 ± 0.18819 [n = 5]	0.9150 ± 0.033431 [n = 5]	1.246 ± 0.23671 [n = 6]	1.090 ± 0.31327 [n = 6]

Notes: All QD dosing and the Day 1 and Day 6 AM dosing of the Q12H administrations occurred under fasted conditions; all other administrations occurred without regard to food intake.

AUCs and C_max values are presented as geometric mean (geometric CV%).

T_max and T_{max, ss} are presented as median (min, max).

One participant in the 3 g QD cohort was excluded from the summary statistics due to multiple consecutive missing blood samples on Day 6.

Other parameters are presented as arithmetic mean ± SD.

Food Effect on Pharmacokinetics

Compared to dosing in the fasted state, the coadministration of a high calorie/high‐fat meal with a single 12 g ribitol dose delayed absorption as evidenced by a change in T_max from approximately 0.5 to 1.5 h and a reduction in C_max by approximately 50%. However, the overall extent of absorption (AUC) was not markedly affected and the 90% confidence interval for the GMR for AUC_∞ and AUC_0–last was between 80% and 125% (Table 3). Because the confidence interval did not include 1, a slight ∼10% decrease in overall exposure with a high‐fat meal cannot be excluded, but an effect of this magnitude would not be considered clinically significant.

Table 3.

Summary of Statistical Comparisons of Baseline‐Adjusted Plasma Ribitol Pharmacokinetic Parameters Following Administration of a Single 12 g Oral Ribitol Dose Under Fed Versus Fasted Conditions

	12 g (Fed) (Test)		12 g (Fasted) (Reference)
Parameter	Geometric LSM	n	Geometric LSM	n	GMR (%)	90% Confidence Interval	Intra‐Subject CV%
AUC0‐∞ (µg h/mL)	402.92	15	451.11	16	89.32	85.08‐93.77	7.46
AUC0‐last (µg h/mL)	401.46	15	449.94	16	89.23	84.94‐93.73	7.55
Cmax (µg/mL)	94.04	15	196.6	16	47.84	41.63‐54.97	22.09

Notes: Parameters were ln‐transformed prior to analysis.

Geometric least‐squares means (LSMs) are calculated by exponentiating the LSMs from the ANOVA.

Geometric Mean Ratio (GMR) = 100*(Test/Reference).

Intra‐subject CV% was calculated as 100 × square root(exp[residual variance]‐1), where MSE = Residual variance from ANOVA.

QT Study Parts 1 and 2

The PK parameters after a 21 g ribitol dose administered in the fasted state in Part 1 versus Part 2 (Table 4) were within the expected variability. However, overall exposure was similar to the highest dose of 15 g which had been administered previously in the SAD study, indicating that an absorption plateau may have been reached. The moxifloxacin exposure after a single 400 mg dose was within that reported in similar studies. ¹¹

Table 4.

Geometric Mean (%CV) Pharmacokinetic Parameters of Ribitol 21 g and Moxifloxacin 400 mg Administered as Single Doses in the Fasted State

Pharmacokinetic Parameters	Ribitol 21 g (Part 1) n = 6	Ribitol 21 g (Part 2) n = 27	Moxifloxacin 400 mg n = 27
Cmax (µg/mL)	203.5 (30.1)	250.7 (23.2)	1.688 (29.4)
AUC∞ (µg h/mL)	538.0 (12.8)	760.2 (23.5)	26.090 (24.6)
AUC0‐last (µg h/mL)	535.5 (12.8)	757.2 (23.5)	24.630 (25.0)

QTc Analyses

By‐Time‐Point (Central Tendency) and Categorical Analyses Results

LS mean change‐from‐baseline QTcF (ΔQTcF) on ribitol closely followed the placebo pattern across post‐dose time points (Figure 2). LS mean placebo‐corrected ΔQTcF (ΔΔQTcF) on ribitol varied from −1.3 ms (at 1 h post‐dose) to 2.2 ms (at 4 h post‐dose). The upper bound of the 90% of ΔΔQTcF was well below 10 ms at all post‐dose time points. Dosing with moxifloxacin 400 mg resulted in the expected clear increase of LS mean ΔΔQTcF with a peak value of 12.8 ms (90% CI: 10.57 to 14.97) at 4 h post‐dose. The lower bound of the 90% of the estimated ΔΔQTcF was above 5 ms at all post‐dose time points between 1.5 and 8 h.

Figure 2.

Change‐from‐baseline QTcF (ΔQTcF) at each time point with statistical modeling. LS mean and 90% CI based on a linear mixed‐effects model: ΔQTcF = Baseline QTcF + Treatment + Time + Time × Treatment + Period + Sequence. An unstructured covariance structure was used to specify the repeated measures (time for participant within treatment period). The horizontal gray dashed line shows the 10 ms threshold.

There were no participants with QTcF > 480 ms or ΔQTcF > 30 ms. There was one participant with QTcF >450 and ≤480 ms in the moxifloxacin treatment period at two time points. There was one participant with treatment‐emergent inverted T waves at one time point in the ribitol period and at one time point in the placebo period. There were no other participants with treatment‐emergent morphology changes or U waves.

Concentration‐QTc Results

Before the concentration‐QTc analysis was performed, the model‐independent checks of assumptions using exploratory plots were explored as described below.

Assumption 1: No drug effect on heart rate. In the by‐time‐point analysis, LS mean ΔΔHR was below 10 bpm across all post‐dose time points for ribitol, thereby demonstrating that ribitol has no clinically relevant effect on heart rate.

Assumption 2: The selected primary QT correction method should be independent of heart rate. Analysis of the QTcF‐RR relationship indicated that the Fridericia formula used adequately corrected the QT interval for heart rate.

Assumption 3: No hysteresis (time delay) between drug concentrations and ΔΔQTcF. Hysteresis loops of plasma concentration showed that ΔΔQTcF varied without relation to the ribitol plasma concentrations, and hysteresis was therefore not deemed to be present.

Assumption 4: Linear concentration‐QTc relationship. A scatter plot of observed ΔQTcF and concentration for ribitol with a locally weighted scatter plot smoothing (i.e., LOESS as described by Cleveland et al) ¹² line and 95% CI with optimal smoothing parameters selected by the AIC with correction (Hurvich et al) ¹³ and a simple linear regression line provided support for the adequacy of the linear mixed‐effects model used.

The relationship between the individually observed plasma concentrations of ribitol and estimated placebo‐adjusted ΔQTcF is shown in Figure 3. The estimated population slope of the concentration‐QTc relationship was −0.0000020 ms (90% CI: −0.00000880 to 0.00000470; P  =  .6174), with a treatment effect‐specific intercept of 0.46 ms per ng/mL (90% CI: −0.331 to 1.255; P  =  .3374). Neither the treatment effect‐specific intercept nor the slope for ribitol plasma concentrations was statistically significant at 10% significance level. The model predicted effects on ΔΔQTcF were −0.05 ms (90% CI: −1.44 to 1.34) at the geometric mean C_max after the ribitol 21 g dose (250.7 µg/mL). An effect on ΔΔQTcF exceeding 10 ms can be excluded within the full observed range of ribitol plasma concentrations up to ∼351.9 µg/mL. As noted above, ribitol did not cause a clinically relevant change in heart rate. The same is true for effects on cardiac conduction as measured by the PR and QRS intervals.

Figure 3.

Scatter plot of estimated placebo‐adjusted ΔQTcF and ribitol plasma concentrations. The solid black line with dotted black lines denotes the model‐predicted mean ΔΔQTcF with 90% CI, which is based on the estimate from a linear mixed‐effects model with ΔQTcF as the dependent variable, time‐matched ribitol plasma concentration as an explanatory variable, centered baseline QTcF as an additional covariate, treatment (active = 1 or placebo = 0) and time as fixed effects, and a random intercept and random slope per participant and calculated from the equation ΔΔQTcF (ms) = 0.46 (ms) − 2E‐6 (ms per ng/mL) × ribitol plasma concentration (ng/mL). The plotted points denote the pairs of observed ribitol plasma concentrations and estimated placebo‐adjusted ΔQTcF by participant for ribitol or placebo group. The individually estimated placebo‐adjusted ΔQTcFi,k equals the individual ΔQTcFi,k for participant i administered with ribitol or placebo at time point k minus the estimation of the time effect at time point k. The horizontal dashed blue line shows the 10 ms threshold.

Discussion

In summary, the first‐in‐human study demonstrates that ribitol was well tolerated when administered as single or multiple oral doses over 6 days to healthy adults. The PK profile was characterized by rapid absorption, with peak plasma concentrations typically reached within 1 h, followed by a multiexponential decline with a terminal half‐life of 9 to 12 h. Exposure increased in a mostly dose‐proportional manner from 0.5 to 15 g, a range that encompasses the 12 g BID therapeutic dose for patients weighing >50 kg and 9 g BID for patients weighing >30 and ≤50 kg. The increase in exposure is less than dose proportional between 15 and 21 g, with 21 g likely representing the maximum achievable exposure. The PK was predictable and steady‐state exposure can be accurately predicted from single‐dose administration.

The FE assessment indicated that while a high‐fat meal delayed absorption and reduced the peak concentration (C_max) of ribitol by approximately 50%, the total systemic exposure (AUC) remained unchanged. Given ribitol's mechanism of action, in which the therapeutic benefit derives from its role as a biochemical substrate and precursor for CDP‐ribitol, the reduction in C_max is unlikely to affect clinical efficacy. In this context, the overall supply of the substrate rather than its transient peak levels is expected to be the more relevant pharmacological parameter. Accordingly, it is concluded that ribitol may be taken without regard to food intake.

A dedicated QT study using a supratherapeutic 21 g ribitol dose demonstrated the absence of an effect of ribitol on QTcF with a flat concentration‐QTcF relationship. A clinically significant effect on ΔΔQTcF exceeding 10 ms at the upper bound of the 90% confidence interval ¹⁰ was excluded across the full range of ribitol plasma concentrations in the study, up to 351.9 µg/mL, the highest level observed. To put this in perspective, the highest expected clinical exposure is 239 µg/mL, corresponding to the 95th percentile in a 30 kg patient receiving ribitol 9 g twice daily. The 95th percentile is considered to be the “high clinical exposure” because there are no extrinsic factors expected to affect ribitol exposure. No drug–drug interactions are expected given the lack of any significant metabolism and the lack of interactions between ribitol and key CYP enzymes and transporters. The only intrinsic factor with the potential for a change in exposure is renal impairment given that 20% of the administered ribitol dose is excreted unchanged in the urine. However, should this be the case as evidenced in a future renal impairment study, a dose adjustment would be implemented to maintain therapeutic exposures. Similarly, the therapeutic dosing regimen—9 g BID for patients weighing >30 to ≤50 kg and 12 g BID for those >50 kg—was designed to achieve a similar therapeutic concentration range across weight groups. The 239 µg/mL value represents the upper bound of this exposure range. Administration of the positive control moxifloxacin 400 mg resulted in the expected positive and statistically significant concentration‐QTcF relationship with a peak effect on ΔΔQTcF of 13.73 ms. Thus, appropriate assay sensitivity was established in this QT study. The absence of a clinically significant effect of ribitol on QTc, as demonstrated in the clinical study, is consistent with the preclinical findings. In vitro screens against 11 cardiac ion channels associated with proarrhythmic risk showed no effect at free plasma concentrations up to 3 mM (441 µg/mL); inhibition of the hERG channel (IKr) occurred at an IC50 of >2.9 mM (426 µg/mL). Also, a dedicated cardiac safety study in male Bama minipigs administered ribitol at doses of 0.3, 1, and 2 g/kg revealed no QTc effect.

The administration of exogenous ribitol is intended to provide a direct biochemical precursor to CDP‐ribitol, leading to increased substrate availability for the glycosylation of αDG. Alternative precursors, such as ribose, have been considered in this context, but their therapeutic potential is constrained by safety concerns. Ribose supplementation at the required doses can cause hypoglycemia and the formation of advanced glycation end products. Moreover, the aldehyde in ribose can interact with proteins and alter their function. ¹⁴ , ¹⁵ For these reasons, health authorities have issued formal safety guidance restricting ribose dosing to a maximum of 36 mg/kg (∼2.5 g for a 70 kg adult), well below the approximately 12 g BID dose level necessary for a meaningful therapeutic impact with ribitol. ¹⁶ In contrast, ribitol has a favorable safety profile and has been well tolerated up to 21 g, the highest dose tested. Neither preclinical nor clinical studies with ribitol have identified hypoglycemia as a risk, even at doses well above ribose safety thresholds. In comparative metabolomics studies, it was also shown that ribitol led to greater ribitol‐5‐ phosphate levels and greater matriglycan expression, a marker for αDG glycosylation, than other pentose alcohols, including ribose. ¹⁷ These observations indicate that ribitol is likely to be a safer and more viable therapeutic option for augmenting αDG glycosylation in patients with FKRP‐related disorders such as LGMDR9.

In designing this study, muscle biopsies were not pursued, as healthy adult participants would be expected to have normal αDG glycosylation and thus would not demonstrate a detectable increase with ribitol administration. Instead, safety, tolerability, and pharmacokinetics were prioritized as the key endpoints appropriate for a first‐in‐human investigation. Future studies in LGMDR9 patients will be required to assess the extent to which ribitol supplementation enhances αDG glycosylation and, ultimately, improves muscle integrity and function.

Taken together, the results of the early clinical evaluations in healthy adults described here support the further clinical development of ribitol as a therapeutic intervention for LGMDR9. The compound was well tolerated across a broad dose range, showed predictable PK, absence of an effect on QTcF, and was not significantly impacted by food with respect to overall bioavailability. The absence of adverse metabolic effects, particularly hypoglycemia, further distinguishes ribitol from related metabolic substrates such as ribose. These findings lay the foundation for subsequent patient‐based trials designed to evaluate whether ribitol can meaningfully increase αDG glycosylation, slow disease progression, and improve outcomes in individuals living with LGMDR9.

Conflicts of Interest

At the time of this analysis all authors were employees or contractors of BridgeBio Pharma Inc. and ML Bio Solutions Inc., the sponsors of the studies. The authors report no other conflict of interest or financial relationship.

Funding

BridgeBio Pharma Inc. and ML Bio Solutions Inc. sponsored this study.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.