Healthy volunteers were recruited through flyers posted on Kanazawa University notice boards. Participants completed a detailed questionnaire that was developed to collect demographic, health, and dietary information.

Eligible participants in the study included both males and females who were aged above 20 years. Individuals were excluded if they had (i) a disease condition, such as dementia or liver, renal, or heart dysfunction; (ii) a history of cancer; (iii) allergies to polyphenols or any drug or food ingredient; or (iv) consumed any supplement containing rosmarinic acid within 15 days of the first administration of M. officinalis extract containing RA. Furthermore, women were excluded if they were pregnant or lactating. A total of eleven healthy individuals aged between 20 and 31 years were enrolled in the two studies [Study 1 (fasted state) and Study 2 (fed state)] using the same inclusion and exclusion criteria. The study statistician generated the random allocation sequence. All subjects were randomly allocated to treatment arms. Eight subjects were male (72.7%); three subjects were female. Seven subjects of the eleven completed both Study 1 and Study 2. Two subjects participated only in Study 1, and the another two subjects joined solely in Study 2. Therefore, nine subjects participated in Study 1, and nine in Study 2. Participants and investigators were masked to treatment allocation throughout the study.

The study was approved by the Kanazawa University Medical Ethics Review Board on February 16, 2011. Written informed consent was obtained from all study subjects upon enrollment. The authors confirm that all ongoing and related trials for this drug are registered. This study is registered at the University Hospital Medical Information Network Clinical Trials Registry (UMIN-CTR: UMIN000004997). The study protocol that was registered was exactly the same as the study protocol approved by the Kanazawa University Medical Ethics Review Board. Subjects were recruited starting in March 2011 and follow-up was completed in June 2011. This study was conducted in accordance to the Declaration of Helsinki in its currently applicable version. A safety monitoring team monitored the progress and data of the trial. The protocol for this trial and supporting CONSORT checklist are available as supporting information; see S1 CONSORT Checklist and S1 and S2 Protocols.

This was a placebo-controlled trial to investigate the safety, tolerability and pharmacokinetics of M. officinalis extract containing RA. To date, the safety dosage of RA has not yet been investigated in humans. Acceptable daily intake (ADI) is a measure of the amount of a specific substance in food that can be orally ingested on daily basis over a lifetime without appreciable health risk [21]. An ADI value can be calculated from the no-observed-adverse-effect level (NOAEL) dividing by one hundred [21]. In our study using AD model transgenic mice (Tg2576), RA exhibited no adverse effect considered drug-related at a NOAEL of 1g/ kg BW/ day in the 10 months study [12]. When human body weight was assumed to be 50 kg, NOAEL was determined 50 g per day. Then, the ADI value of RA was calculated 500 mg per day.

The two studies were designed as follows.

Study 1 (fasted state): Each participant received three separate interventions (investigation #1, #2, and #3) designed by groups A, B, or C in a 3 × 3 Latin square design with the order of consumption randomized and separated by a ten-day interval (wash-out period) (Fig 1). In details participants who allocated to group A were applied placebo in the investigation #1, and were applied M. officinalis extract containing RA 250 mg in the investigation #2 and were applied M. officinalis extract containing RA 100 mg in the investigation #3. Similarly, group B was applied M. officinalis extract containing RA 250 mg, placebo and M. officinalis extract containing RA 500 mg in the investigations #1, #2, and #3 respectively. Group C was applied M. officinalis extract containing RA 100 mg, M. officinalis extract containing RA 500 mg and placebo in these investigations #1, #2, and #3 in order given (Fig 1).

Each intervention was conducted for 10 days, consisting of a 7-day controlled diet period and a 3-day intervention period. In order to avoid the effects of polyphenols derived from other food sources during the intervention and controlled diet periods, participants were asked to consume foods that contain minimal amounts of polyphenols, such as rice and fish. The participants fasted for 12 h before the administration of M. officinalis extract containing RA or placebo.

Study 2 (fed state): In this fed state arm, placebo or M. officinalis extract containing RA was administered 30 min after completion of the standardized low-polyphenol meal, i.e. rice ball and clear soup. The other procedures of Study 2 were carried out in the same procedure as in the Study 1.

The primary outcome was to investigate pharmacokinetics of M. officinalis extract containing RA including food-effect in human. The secondary outcomes were to evaluate the safety and tolerability of M. officinalis extract containing RA.

Placebo and RA capsules (M. officinalis extract containing RA) were provided free from Maruzen Pharmaceuticals Co., Ltd., Japan. Lemon balm leaves (M. officinalis) were extracted with hot water and were purified with ethanol to obtain rosmarinic acid-rich extract. Each RA capsule contained 185 mg M. officinalis extract along with 50 mg RA, 60 mg lactose, and 5 mg calcium sterate. Each placebo capsule contained 210 mg lactose, 35 mg caramel, and 5 mg calcium stearate. There was no detectable RA content in the placebo capsule. Placebo capsules were made of the same size and color as the RA capsules. Because each RA capsule contained 50 mg RA, participants in the RA 100 mg, 250 mg, and 500 mg administration groups were allocated 2, 5, or 10 RA capsules, respectively. Participants in the placebo group were administered 2 placebo capsules.

For the purpose of the trial, physicians interviewed the participants, and performed clinical and physical examinations during their visits. They assessed the subjects’ physical condition and subjective symptoms and determined whether any adverse events were experienced. The health status of each subject was examined during physical examination. The examination parameters for the trial included height, weight, body mass index, and blood pressure, which were recorded from the subjects’ brachial region while they were seated.

Laboratory blood chemistry parameters, including a hematology (white blood cell count, red blood cell count, hemoglobin levels, and hematocrit and platelet count) and blood biochemistry [levels of aspirate aminotransferase (AST), alanine aminotransferase (ALT), lactate dehydrogenase (LDH), γ-glutamyltransferase (γ-GTP), total bilirubin (T-Bil), blood urea nitrogen (BUN), and creatinine (Cr)], were assessed with blood samples obtained at baseline and 48 h after the intake of M. officinalis extract containing RA or placebo. Baseline blood samples were collected from the intermediate cubital vein after the subjects had fasted for 12 h, followed by further samples being collected 0.25, 0.5, 1, 2, 3, 6, 24, 48 h after intake of M. officinalis extract containing RA or placebo. All the samples were processed and analyzed in a laboratory (SRL. Inc, Tokyo, Japan).

In the course of the clinical trial, adverse events associated with the study agents were self-reported by the participants, and blood analyses including blood cell counts and blood chemical parameters were monitored. All observed and self-reported adverse events, regardless of suspected causal relationship to the study treatments, were recorded on the adverse event form throughout the study.

Serum RA concentration is measured by High-Performance Liquid Chromatography-electrochemical detector analysis using an ESA coulometric detection system (ESA Inc., Boston, USA). Quantitative determination of RA is performed by using an external standard method [22], which verified that the detector response is linear for a concentration of up to 500 μmol/L for RA. In brief, 50 μL of serum were added 50 μL of 0.1 mol/L sodium acetate buffer (pH 5.0) and 200 μL of 0.83 mol/L acetic acid in methanol. The mixture was vortexed, sonicated, and centrifuged (at 8500g for 5 min at 4°C), and the supernatant was injected onto a High-Performance Liquid Chromatography-electrochemical detector C18 column (ODS150, MC Medical, Inc., Tokyo, Japan). The mobile phase A (solvent A) was 50 mmol/L sodium acetate containing 5% methanol (pH 3.0 adjusted with phosphoric acid), while the mobile phase B (solvent B) was 50 mmol/L sodium acetate containing 40% acetonitrile and 20% methanol (pH 3.5 adjusted with phosphoric acid). We used the following elution profile (0.6 mL/min): 0–28.5 min, linear gradient from 85% solvent A/15% solvent B to 20% solvent A/80% solvent B; 28.5–31 min, isocratic elution with 0% solvent A/100% solvent B; and 31–35 min, isocratic elution with 85% solvent A/15% solvent B. The eight electrode detector potentials were increased from 0 to 700 mV in increments of 100 mV. It was reported that the majority of RA were present in the blood as conjugated forms [20]. The difference in RA content before and after glucuronidase treatment was assumed to be attributed to the amount of the sulfate and glucuronide conjugates in the sample. Serum (50 μL) was mixed with 50 μL of glucuronidase type H-5 (Sigma-Aldrich. Inc. St. Louis. MO. USA) solution in 0.1 mol/L acetate buffer (pH 5.0) containing both 1.3 units of sulfatase and about 200 units of β-glucuronidase activity. The mixture was incubated at 37°C for 45 min.

We chose the sample size to assess pharmacokinetics of M. officinalis extract containing RA in six subjects in each treatment arm. The one-way analysis of variance test was used to determine the significance of any differences among groups. The Chi-square test was used to determine the significance of gender. In the safety evaluation, defined as the blood analysis performed for each treatment arm, two-ways repeated measures analysis of variance was used to assess measurements. The SPSS software package (version 12.0J; SPSS Inc., Chicago, IL, USA) was used to perform statistical analyses. Area under the curve (AUC) for RA was calculated using the PK functions plug-in for Microsoft Excel [23]. The maximum serum concentration (C_max), and time at which the maximum serum concentration was observed (T_max) were determined from individual serum pharmacokinetic curves. We calculated AUC, C_max, T_max for RA total, intact form and conjugate form, separately. The residual error was used to calculate the 90% confidence interval (CI) for the log-state fed-fasted difference. If the anti-logs of the estimate 90% CI of fed-fasted difference were within the limits of 0.80–1.25 and 0.70–1.43 for AUCs and C_max, respectively, no food effect would be claimed [24]. Data were expressed as mean (SEM). A two-sided p-value of <0.05 was considered statistically significant in all the analyses.

Baseline characteristics were collected from all eleven eligible participants. With the exception of one, none of the participants withdrew before the end of the study; the withdrawing participant in the M. officinalis extract containing 100mg RA arm in fasted state dropped out halfway through the study because of illness induced by insufficient sleep. Nevertheless, the safety monitoring team speculated that this incident was unlikely to have been related to the study protocol. Because no adverse effects were recorded in the other intervention periods in which he took M. officinalis extract containing 500 mg RA and placebo in fasted state, and M. officinalis extract containing 100 mg, 500 mg RA and placebo in fed state. In the remaining participants, baseline characteristics were similar among all the different treatment groups (Table 1).

In the past history, one subject of group A had retinal detachment, one subject of group B mediastinal emphysema, another subject of group B dyslipidemia and renal calculus, and one subject of group C a history of surgery for endocardial deficiency. No subject had a habit of smoking. Two subjects of group B had a habit of drinking 1 or 2 bottles of beer per week. The other subjects had no habit of drinking.

Blood cell count data of one subject in group B were not available for coagulation. No significant changes were observed in the blood parameters between treatment arms (placebo, M. officinalis extract containing RA 100mg, 250mg or RA 500 mg) (Table 2).

The serum levels of AST, ALT, LDH, γ-GTP, and T-Bil (indicators of liver function) or those of BUN and Cr (indicators of kidney function) were not significantly affected by any of the interventions (M. officinalis extract containing 100 mg, 250 mg, 500 mg RA, or placebo) (Table 2).

In the placebo group, no RA was detected in the serum in all subjects neither in fasted nor fed state (data not shown). After administration of M. officinalis extract containing 100 mg RA, no RA was detected in serum samples from 4 of 6 subjects in fasted state or from 5 of 6 subjects in fed state (data not shown). The mean serum concentration-time profiles of RA after administration of a single dose of M. officinalis extract containing RA (500mg fasted/ fed state and 250mg fasted state) are shown in Fig 2.

Maximum concentration of total RA for the fasted regimens was reached 1h after oral administration of a single M. officinalis extract containing 250 mg RA and was 72.22 ± 12.01 nmol/ L (mean ± SEM), whereas no RA was detected in serum samples from 5 of 6 subjects in fed state. After administration of M. officinalis extract containing 500 mg of RA, total RA reached a maximum level of 162.20 ± 40.20 nmol/ L 1h and 142.20 ± 45.20 nmol/ L 3h, in fasted and fed state, respectively.

The noncompartmental pharmacokinetics analyses of intact and conjugated RA after administration of a single dose of M. officinalis extract containing RA (500 mg fasted/ fed state or 250 mg fasted state) are shown in Table 3.

When we administrated M. officinalis extract containing 500 mg RA in fasted state, the mean value of AUC_Total was increased to 2.0 times the AUC_Total of M. officinalis extract containing 250 mg RA administration in fasted state. Especially, the mean value of AUC_Intact of M. officinalis extract containing 500 mg RA fasted state was increased to 7.3 times the AUC_Intact of M. officinalis extract containing 250 mg RA fasted state. Median T_{max Total} values were similar in M. officinalis extract containing 250 mg and 500 mg RA fasted state. The mean value of C_{max Total} and C_{max Intact} of M. officinalis extract containing 500 mg RA fasted state were increased to 2.0 and 4.0 times the C_{max Total} and C_{max Intact} of M. officinalis extract containing 250 mg RA fasted state respectively (Table 3). Three subjects (group B) were included in both treatment arms [M. officinalis extract containing 250 mg RA (fasted) and 500 mg RA (fasted)], we showed the effect size along with its 95% CI (Table 4).

Regarding food-effect evaluation, significant increase in the mean value of AUC_Total (1.3 times) was observed with food intake. Statistical analysis result demonstrates that 90% CIs of AUC_Total ratio of M. officinalis extract containing 500 mg RA fed state to fasted state were significantly below the lower limit (0.80) and were significantly above the upper limit (1.25) (Table 5).