The trial was conducted in compliance with the International Conference on Harmonization (ICH) guidelines and the ethical principles of the Declaration of Helsinki and its subsequent amendments. The trial design was approved by the IEC/IRB of North Rhine Medical Council (IRB number: 2018188). All trial procedures were conducted by Profil Institute for Metabolic Research GmbH (Neuss, Germany) in collaboration with MLM Medical Labs GmbH (Moenchengladbach, Germany). The trial is registered at clinicaltrial.gov (NCT05041179).

The study flowchart is shown in Figure 1. 729 participants were voluntarily recruited from the database and a written informed consent was obtained from 217 subjects during an ambulatory visit (Visit 0), where eligibility and willingness to participate were assessed. Inclusion criteria for eligibility were: age range 20–40 years old (young adults) and 60–85 years old (older adults), both inclusive; both sexes were admitted; BMI range 18.5–30 kg/m² (young adults) and 25–35 kg/m² (older adults), both inclusive; considered generally healthy; sedentary for older adults, with less than 1 h of strenuous exercise per week; levels of HbA1c less than 5.7% (38.8 nmol/mol for young adults) and 6.5% (47.5 nmol/mol for older adults). Subjects with known clinically relevant comorbidity, acute illness, serious infectious disease for 4 weeks prior the product intake, abnormal screening laboratory tests, and abnormal blood pressure (systolic blood pressure <90 mmHg or >139 mmHg and/or diastolic blood pressure <50 mmHg or >89 mmHg) were excluded, although subjects with normal blood pressure taking anti-hypertensive medication were included. Heavy smokers (more than five cigarettes per day) and people with history of alcoholism or drug abuse were also excluded and participants were asked to refrain smoking 3 days prior and during the intervention. Participants were also asked to avoid any medication for 14 days before test product intake, consumption of high protein supplements within 60 days of screening and during the study, and consumption of any antioxidant, vitamins, herbal supplements within 2 weeks prior to and during the study. The full list of exclusion criteria is available in Supplementary Table S2. Conformity of participants to all exclusion criteria was assessed during the screening visit (Visit 1), during which a clinical examination and a blood sample was taken; a total of 137 participants were selected to enter the study (20 young adults, non-interventional group; 117 older adults, interventional group).

The study is a single-center, double-blind, randomized, placebo-controlled 4-arm study design. Following the screening visit, the interventional group of older adults have undergone three visits on day 1 (Visit 2), day 4 (Visit 3), and day 15 (Visit 4) of the treatment supplementation. At Visit 2, participants were randomized in a 1:1:1:1 ratio in four different arms: placebo, 2.4 g of actives, 4.8 g of actives, 7.2 g of actives. Participants came to each visit after an overnight fast (>10 h), with up to 200 ml of water allowed, avoiding strenuous exercise within the last 48 h, avoiding a protein-rich meal in the evening before the visit as well as consumption of alcohol, caffeine and/or methylxanthine-containing products in the last 24 h before the visits. At each visit the subjects had a medical examination (body measurements, vital signs, and physical examination). Two blood samples were taken 60 min before and 60 min after the product intake for assessment of glycine, total cysteine, GSH-T, GSH-F, GSH-F:GSSG in whole blood and glycine, total cysteine, cystine, MDA in plasma. Subjects received a mixed meal (breakfast) after the product intake.

A powder blend of N-acetylcysteine (NAC) and glycine was tested at three different daily doses: 2.4 g of actives (1.2 g NAC + 1.2 g glycine), 4.8 g of actives (2.4 g NAC + 2.4 g glycine), and 7.2 g of actives (3.6 g NAC + 3.6 g glycine) per day. Each dosage was split into two doses consumed in the morning and in the evening. The placebo control arm received 7.2 g of isomaltulose with the same way of administration. In order to keep it blind, all subjects took three sachets of test product at each dosing; all sachets had the same weight of 1.2 g. The scheme of the sachet composition for each arm is shown in Supplementary Table S1. The investigational product was manufactured, filled, and labelled by Glanbia Nutritionals GMBH, Germany. Secondary and tertiary packing of the study product for blinding purposes was performed at Profil Institute.

GSH-T, GSH-F, GSH-F:GSSG, total cysteine and glycine levels were measured in whole blood. After blood withdrawal into EDTA, samples were rapidly stored at −80°C. The quantitative analysis of free GSH and GSH-T was performed according the applicable international standards from the FDA and EMA, using the HPLC method from Chromsystems (Gräfelfing/München, Germany). Protein precipitation was performed on all samples, standards and controls. Prior to analysis, samples underwent a protein precipitation step followed by a derivatization step. To detect both forms of glutathione (GSH-F and GSSG), each sample is divided into two aliquots after the initial protein precipitation. For detection of GSH-F, the first aliquot is derivatized immediately. The second aliquot is first chemically reduced before derivatization, which leads to the conversion of GSSG to GSH. In this second aliquot the sum of both oxidized and reduced glutathione (GSH-T) can be analyzed. The subsequent analysis bases on HPLC determination with fluorescence detection. The concentration of the oxidized form GSSG was calculated from the obtained results using the formula GSSG = GSH-T - GSH-F x 0.99672. A similar method was used for the MDA quantification, assessed also by HPLC Chromsystems (Gräfelfing/München, Germany). The analysis was performed using the Chromsystem reagent kit, that allows the determination of MDA on an isocratic HPLC system with fluorescence detection. Before the fluorescence detection, EDTA plasma samples and controls were prepared with a precipitation step followed by derivatization.

Quantification of glycine, total cysteine, cystine and NAC in plasma was performed with Ultra Performance Liquid Chromatography tandem Mass Spectrometry analysis (UPLC-MS/MS). Samples, standards and controls were added with internal standards working solution and tris (2-carboxyehtyl) phosphine hydrochloride solution (Merck KGaA, Darmstadt, Germany), or replace by Milli-Q water for cystine measurements, and vortexed at 800rpm for 15 min at room temperature. Protein precipitation using methanol (Merck KGaA, Darmstadt, Germany) and 0.1% formic acid (Merck KGaA, Darmstadt, Germany) was performed and the collected supernatant was treated with a derivatization solution of aminoquinolyl-N-hydroxysuccinimidyl carbamate (AQC) (Merck KGaA, Darmstadt, Germany). After incubation at 55°C, a final dilution of the derivatized mixture was done by an addition of ammonium formate solution 0.55 g/L (Merck KGaA, Darmstadt, Germany) in water containing 0.1% formic acid (Merck KGaA, Darmstadt, Germany). The samples were then injected into the UPLC-MS/MS for analysis. Chromatographic separation for cystine measurements was performed using AccQ-TAG Ultra C18 100 × 2.1 1.7 um analytical column (Waters^®, United States), for glycine, cysteine, and NAC using an UPLC YMC-Triart metal Free C18 50 × 2.1 1.9 um analytical column (YMC^®, Japan). Mass spectrometric analysis and detection were performed using specific MRM for analytes and their related labeled internal standards, on a Xevo TQ-XS (Waters^®, United States), equipped with electrospray ionization source in positive mode and argon as collision gas. Integration of chromatographic peaks was performed using Masslynx software (Waters^®, United States). The analytical method for the quantification of glycine in whole blood is based on precipitation, centrifugation and dilution of the supernatant, followed by the determination of the glycine using Liquid Chromatography tandem Mass Spectrometry analysis (LC-MS/MS). Samples, standards and quality controls were added with internal standards working solution and precipitated with acetonitrile (Chemsolute, Th.Geyer, Renningen, Germany) and 0.1% formic acid (Merck KGaA, Darmstadt, Germany). The supernatant was transferred in a 96-well plate and submit to LC-MS/MS analysis using Mass Spectrometer TSQ Vantage (Thermo Fisher Scientific, Waltham, Massachusetts, United States). The time point of 60 min to determine oral bioavailability of the interventional product was decided based on the published t_max of approximately 45 min after oral ingestion of NAC (Borgstrom and Kagedal, 1990) and that of approximately 75 min of glycine in blood plasma of healthy volunteers (Butterworth et al., 1958).

Safety of the product was assessed during the study through the monitoring of adverse events and blood biochemistry analysis performed by MLM Medical Labs. Basal blood hematology and biochemistry were performed at the screening Visit 1 on all subjects. At the end of the study at Visit 4, a short list of hematology and biochemistry markers was performed to monitor the safety of the product intake at−60 min before and 120 min after the product intake. The parameters assessed were: hematocrit, hemoglobin, erythrocytes, mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH), MCH concentration, platelets, leucocytes, sodium, potassium creatinine, glucose, total cholesterol, aspartate aminotransferase (GOT), alkaline phosphatase, gamma-glytamyltransferase (ɣ-GT), insulin, triglycerides. An exhaustive list of all parameters detected can be find in Annex 1.

The sample size calculations for the interventional cohort are based on unpublished data from Profil which evaluated total GSH measurement in healthy young subjects. A sample size of 120 subjects (4 arms; 30 per arm) would provide 80% power to detect a significant intraindividual change in total GSH of 12.5% (comparison baseline to end-of-treatment) and significant interindividual difference between groups (0 vs 2.4 vs. 4.8 vs. 7.2 g) of 17.5% at α = 0.05 level. To account for potential outliers, an additional eight subjects (2 per arm) results in a total sample size of N = 128. For the non-interventional cohort, a group size of 20 subjects will detect a minimum difference in baseline GSH of at least 16% between young and elderly healthy subjects (0.05 alpha, 0.8 power) based on an inter-subject CV of 22%. Demographic and general baseline characteristics are represented as LS-means with standard deviations for normally distributed parameters or medians with IQR for non-normally distributed parameters. The baseline values of the young adults were compared to the baseline parameters of older adults using parametric t-statistics for LS-means and one-tailed nonparametric Wilcoxon/Mann-Whitney tests for medians. Correlation analysis between baseline parameters were performed in older adults with all arms pooled. A linear mixed model was used to test intraindividual changes from baseline (Visit 2) and end of study (Visit 4) within each dose group and from placebo group. All hypothesis tests were assessed at 5% level.

Data related to this study as well as a clinical study report released by Profil and Nestlé Health Science are available upon reasonable request.

117 older adults between 60 and 85 years of age were enrolled and randomized into a placebo group and three groups receiving active doses at 1.2, 2.4, and 3.6 g of each of Gly and NAC daily (Figure 1; Supplementary Table S1). 114 of the participants completed 2 weeks of daily GlyNAC administration orally taken in two doses (Figure 1). The cohort consisted of generally healthy older adults: subjects were excluded if they had hypertension (>139/89 mmHg), if they had a fasting glycated hemoglobin A1c (HbA_1c) > 6.5% (47.5 nmol/mol), indicative of diabetes, a diagnosis of diabetes, a BMI >35 kg/m², or if they suffered from other major chronic medical conditions, such as dementia, frailty, or malignancies (Supplementary Table S2). A non-interventional control cohort of 20 healthy adults between 20 and 40 years of age was recruited as a reference group. Baseline characteristics of the non-interventional cohort of young adults and the interventional cohort of older adult volunteers are summarized in Table 1. Several markers related to metabolic health, including HbA_1c, triglycerides, LDL, total cholesterol, and C-reactive protein (CRP) were moderately increased in healthy older adults. The mean BMI of 28 kg/m² is in the range of overweight category but is within the normal distribution in a healthy cohort of older adults (Peng et al., 2019). Consistent with a moderate increase of these well-described changes in healthy aging, we observed significantly higher levels of malondialdehyde (MDA), a marker for circulating levels of lipid peroxides, in older adults compared to the levels of young subjects (means, young = 0.136, older adults = 0.158 µmol/L, p-value < 0.0001) (Table 1; Figure 2A). We also observed higher levels of total cysteine in plasma (means, young = 276, older adults = 314.8 µM, p-value < 0.0001) (Figure 2B), representing the pool of reduced cysteine, cysteine disulfides and protein-bound cysteine after addition of the reducing agent dithiothreitol (DTT) (Fu et al., 2019). Together with these changes, we identified significantly lower levels of circulating glycine in blood plasma (means, young = 271.6, older adults = 229.4 µM, p-value = 0.01) (Figure 2C). We next analyzed the status of the glutathione redox system in whole blood. The levels of GSH-T representing the combined pool of reduced and oxidized glutathione as well as protein-bound glutathione were unaffected in older adults compared to that of young subjects (Figure 2D). However, despite normal levels of GSH-T, the oxidized glutathione (GSSG) levels were significantly increased leading to a lower GSH-F:GSSG ratio in older compared to young adults (Table 1; Figures 2E,F). While GSH-F:GSSG and GSH-T levels were tightly correlated, the oxidative damage markers assessed in this study did not correlate with glutathione status (Supplementary Figure S1). These data show that older age in an overall healthy cohort is associated with higher levels of oxidative stress markers indicating a shift towards a pro-oxidative redox balance.

No severe adverse effects were reported during the course of the study and none of the study subjects discontinued GlyNAC because of adverse effects. Two of the three subjects that discontinued the study opted to not participate in Visit 4 because of COVID-19 recommendations to reduce unnecessary travels. One subject discontinued because of high blood pressure. A comprehensive safety panel of relevant blood markers was quantified from each study subject at the beginning of the study, at 3 days after first dosing and after 2 weeks of GlyNAC consumption, including blood pressure, glucose, insulin, creatinine, Alkaline Phosphatase (ALP), Alanine Aminotransferase (ALT), among others (Supplementary Table S3; Supplementary Data). No significant effects could be observed for any of the markers except for clinically not relevant, but significant changes in ALP (placebo = 60.67; 2.4 g = 65.26, p-value = 0.216; 3.6 g = 70.66, p-value = 0.009; 7.2 g = 70.52, p-value = 0.01, U/L, compared to placebo). In conclusion, 2 weeks of daily GlyNAC consumption is safe and well tolerated.

To compare bioavailability of GlyNAC in blood plasma in response to each dose, we quantified fasting levels of glycine, NAC and oxidized cystine 60 min before and after oral intake. Quantifying the pool of oxidized cysteine can be used as a surrogate endpoint for NAC dosing; acetyl groups are removed from NAC in the gastrointestinal tract and the liver, releasing cysteine into the circulation where it is present in a redox equilibrium between the reduced form, cysteine, and the oxidized form, cystine (Borgstrom and Kagedal, 1990). Measuring oxidized cysteine has the advantage of a higher sensitivity and better practicality in clinical trials using larger cohort sizes to determine the dosing efficacy compared to the direct measurement of reduced cysteine, which is difficult to measure and that can be prone to oxidation artefacts. GlyNAC supplementation led to a dose-dependent increase of circulating levels of glycine in blood plasma measured at 60 min after dose intake at all three doses (Figure 3A). We also detected dose-dependent increases of oxidized cysteine in response to GlyNAC administration (Figure 3B) while total cysteine disulfide levels, which include also protein-bound cysteine, remained unchanged (Supplementary Figure S2A). Thus, oral intake of GlyNAC leads to an efficient and dose-dependent absorption of glycine and NAC within 60 min. In contrast, free NAC is only detectable at the medium and high doses in a subset of subjects (Supplementary Figure S2B). Taken together, glycine and NAC are readily absorbed after oral intake within 1 h at all three active doses.

The primary endpoint of the study was defined as the difference in GSH-T normalized to hematocrit compared to placebo at the end of the study. No significant differences were observed at any of the three doses tested compared to the placebo treated group (LS-means, placebo = 903.5; 2.4 g = 947.7; 4.8 g = 907.7; 7.2 g = 959.6 mg/L per % Ht, p-value 2.4 g vs. placebo = 0.9039; p-value 4.8 g vs. placebo = 0.9340; p-value 7.2 g vs. placebo = 0.2778) and no significant intra-individual increases were detected comparing start of the study to the end of the study (Figure 4A; Table 2). Similarly, GlyNAC supplementation also did not change significantly the ratio of GSH-F:GSSG (LS-means, placebo = 12.49; 2.4 g = 13.53; 4.8 g = 11.34; 7.2 g = 12.65 mg/L per % Ht, p-value 2.4 g vs. placebo = 0.276; p-value 4.8 g vs. placebo = 0.502; p-value 7.2 g vs. placebo = 0.739) (Table 2). Levels of MDA or total cysteine were not impacted by GlyNAC supplementation compared to placebo, although a moderate but significant intra-individual lowering of total cysteine levels was observed at the highest dose (means, 7.2 g, start of study = 415.9, end of study = 396.7 µmol/L, p-value = 0.028) (Supplementary Table S4).

We next asked the question whether glutathione biosynthesis from dietary glycine and cysteine is regulated by cellular glutathione demand. To this end, we performed a post hoc analysis on a subset of the study population that is characterized by increased oxidative stress, as determined by the values above the median for MDA, while showing low levels of baseline glutathione levels, as characterized by values below the median. Subjects from the placebo group did not show differences in total glutathione levels when comparing start with end of the study (medians, V2 = 824.4, V4 = 780 mg/L per % Ht, p-value = 0.734) (Figure 4B). In contrast, pooled analysis of subjects from the 4.8 and 7.2 g treatment groups showed a 10.47% increase of total glutathione levels at the end of the 2 weeks of GlyNAC supplementation (medians, V2 = 819.7, V4 = 905.4 mg/L per %Ht, p-value = 0.016) (Figure 4C). The same analysis but combining the subjects from the 4.8 and 7.2 g groups that are not characterized by high oxidative stress did not show changes in GSH-T (medians, V2 = 947.2, V4 = 920.7 mg/L per %Ht, p-value = 0.516) (Figure 4D). An analysis based on elevated total cysteine (above the median) and low GSH-T (below the median) showed a similar result on GlyNAC efficacy in this subpopulation (medians, V2 = 817.7, V4 = 859.3 mg/L per %Ht, p-value = 0.015) (Supplementary Figure S3). These data suggest that healthy older adults with GSH-T levels that are similar to those of young adults may not utilize intra-cellular GlyNAC to generate additional GSH-T, while a subset of subjects with higher glutathione demand may be responding.

An important observation of this study is the lower levels of glycine in healthy older adults (by approximately 15.5% in plasma and 8.9% in whole blood) compared to young subjects (Table 1). We therefore asked the question whether fasting glycine status is related to markers of glutathione status or oxidative stress. However, no significant associations of circulating glycine concentrations with levels of GSH-T, GSH-F:GSSG, or MDA were detected (Supplementary Figures S4A−C). A significant but moderate association of lower levels of glycine with higher levels of total cysteine could be detected (Supplementary Figure S4D). In contrast, older adults from this cohort showed a negative correlation between glycine status and the Homeostatic Model Assessment for Insulin Resistance (HOMA-IR), a composite score indicating differences in insulin sensitivity (Supplementary Figure S4E). This observation in older adults is consistent with previous studies that report associations of low glycine status with elevated insulin resistance (Newgard et al., 2009; Wittemans et al., 2019).

We next asked whether 2 weeks of daily supplementation with GlyNAC increased fasting levels of glycine towards that of young subjects. To this end, we compared glycine levels at the end of the study between the four treatment groups while adjusting for baseline glycine concentrations. We excluded the group above the third quartile of glycine levels to remove subjects that may not have adhered to the overnight fast or that have particularly high fasting glycine levels due to other reasons. We found that the medium and high doses efficiently increased glycine levels compared to baseline (4.8 g coefficient = 65.10, p-value = 0.021; 7.2 g coefficient = 74.9, p-value = 0.007, linear model ANCOVA, reference level: placebo), whereas the 2.4 g groups did not show a significant effect (2.4 g, coefficient = 29.99, p-value = 0.27) (Figure 5; Supplementary Tables S5, S6). Of note, the medium and high doses were able to increase glycine levels close to the median level observed in young individuals (Figure 5, red dashed line). Changes in blood plasma in response to GlyNAC supplementation were not significant, although the medium dose showed a trend towards improved glycine status (4.8 g, p-value = 0.068) (Supplementary Figure S5). In summary, these results suggest that lower fasting glycine values in older adults may be efficiently restored towards intracellular blood levels of young adults within 2 weeks of supplementation.