Method validation
Selectivity and specificity
The ability to accurately measure an analyte in a sample matrix is known as selectivity. Under MRM conditions, the compounds with different m/z will be eliminated before reaching the detector. In human milk, 3-FL has the same molecular weight and MRM transition to 2’-FL. However, the 3-FL was eluted at 2.8 min while the retention time of 2’-FL was 8.1 minutes (Fig. 1), enabling the identification of two linkage isomers.
The standard addition method was employed to verify selectivity and specificity. A comparison was made between peak intensities in the sample, standard, and spiked standard sample. The chromatogram exhibited clear peaks at the same retention time (2.8 min) for the 3-FL standard, human milk, and spiked human milk samples. Additionally, an increase in the intensity of the 3-FL peak was observed in the spiked sample when compared to the human milk sample (see Fig. 1). There is no interfering peak detected at the same retention time as 3-FL.
Linearity and dynamic range
A range of standard concentrations spanning from 2 to 5000 ng/mL was used to evaluate the linearity between the concentration and peak area and to determine the dynamic range of the standard curve. After linear regression, the calibration curve was accepted when the coefficient of determination was greater than 0.995.
Two linear regression equations were estimated to measure a broad range of 3-FL concentrations, covering the ranges of 2 ̶ 500 and 500 ̶ 5000 ppb, independently (refer to Table 1). The coefficient of determination of the calibration curves of low- and high-range concentrations were 0.9999 and 0.9989, respectively (Supplementary Figs. 1 and 2). The measured concentration was backward calculated from the standard curve (Supplementary Tables 2 and 3). The measured concentrations of calibrators were found to be within ± 15% of the nominal concentrations.
The limit of detection (LOD) and quantitation (LOQ)
The limit of detection (LOD) and quantitation (LOQ) were obtained by replicate measurements (n = 7) of the standard at 10 ng/mL. The LOD and LOQ were obtained from the equation (LOD = 3.3x σ / S) and (LOQ = 10x σ / S), respectively. Where S is the slope of the calibration curve, and σ is the standard deviation of the 7 replicates. The limit of detection (LOD) and quantitation (LOQ) were 1 and 5 ppb, respectively (Supplementary Table 4).
Precision and accuracy
The precision and accuracy of the analytical method were assessed. Precision was determined by calculating the coefficient of variation (CV %) from six replicates. The accuracy was measured by the relative error (RE %) between the measured concentration and the nominal concentration, described as a percentage of the nominal concentration. The results were considered to be acceptable when the CV% and RE% were less than 15%.
The precision and accuracy of the analytical method were evaluated both intraday and interday by repeated experiments performed on three separate days using low, medium, and high concentrations of 3-FL (5, 1, and 0.1 ppm). The precisions of intraday were ranging from 0.3–2.3% and no significant difference was found between concentrations. The interday precisions of independent analysis performed on three separate days were slightly higher than intraday precision ranging from 1.0–6.2% (Supplementary Table 5). The accuracy (RE %) of the analysis between different days (intraday accuracy) was less than 10%, and the interday accuracy of 18 data points was less than 5% at all tested concentrations (see Supplementary Table 5). These findings suggest that the analytical method met the precision and accuracy criteria.
Matrix effect
The matrix effect has been validated by the standard addition method. Two different concentrations of 3-FL were spiked into the human milk and compared the measured concentrations obtained from the standard injection and the spiked sample injection.
The results suggest that there seemed to be a slight suppression by the matrix. The percent variations of the 3-FL concentration in the spiked milk sample versus the neat standard solution exhibited negative values ranging from − 8.2 to -8.8% (Supplementary Table 6). The matrix effect could be concluded as not significant based on the fact that no significant differences between replicated samples (SD < 2.2%) suggest the reproducible impact by matrix and the suppression did not seriously increase the LOD or LOQ (< 10%).
Recovery and reproducibility
A series of liquid-liquid extraction (LLE) and solid phase extraction (SPE) has been applied for the sample preparation resulting in a 100-fold dilution of human milk without the chemical reduction step or enzyme reaction. To determine the recovery during the sample preparation, human milk aliquots were fortified with a reduced standard of 3-FL at 25 and 250 ppb concentrations before extraction. Then, the reduced standard were purified in the same way as the blank matrix. The recovery was calculated as the percentage of the 3-FL concentration in extracted samples and the initial added values. Finally, the reproducibility of the extraction protocol was validated by the analysis of three replicates.
The sample preparation method had a high recovery which ranged from 92 to 102% (Supplementary Table 7). The average CV (%) was less than 3% indicating the high reproducibility of the assay. The data indicated that this method is suitable for analyzing 3-FL in human milk.
Sample stability
The stability of human milk samples was assessed under two storage conditions: short-term storage in the autosampler at 4°C for 24 hours and long-term storage at -60°C for four weeks. Two levels of 3-FL were tested under each storage condition with three replicates per level. The results revealed that the CV (%) was less than ± 10% under both storage conditions (Supplementary Table 8). These findings demonstrate that the samples remained stable under the tested conditions.
Quantitation of 3-FL in Korean maternal milk
The 3-FL concentrations in maternal milk were measured from 102 of Korean human milk samples. The concentrations of 3-FL in human milk of secretors/non-secretors during lactation are shown in Table 1. The secretor and non-secretor status of mother were determined by the concentration of 2’-FL in maternal milk (Nguyen, et al., 2022).
The average concentration of 3-FL in Korean mother's milk was 1.6 ± 1.2 g/L (n=102), with a median concentration of 1.4 g/L. In comparison, a previous study conducted in Germany found that the 3-FL level was only 0.28-0.67 g/L during the first 100 days of postpartum (Csernak, et al., 2020). The concentration of 3-FL in the United State was less than 0.7 g/L for the first 6 months (Plows, et al., 2021). These results indicate that Korean mother's milk had a quite high concentration of 3-FL.
The average concentration obtained from the mothers of the first two months of breastfeeding was 1.0-1.1 g/L (n= 26). Measwhile, the average concentration of maternal milk from the later than the third month of lactation mothers was higher as 1.8 g/L (n=76). It was found in the previous studies, 3-FL concentration in United States mothers increased from 0.7 g/L to 2-3 g/L during the first 6 to 24 months of lactation (Plows, et al., 2021).
The results were intriguing that the concentration of 3-FL in non-secretors was higher than in secretors. The comparison of 3-FL concentration between same lactation period exhibitied the same trait as well. For example, the 3-FL concentration in non-secretor mother's milk of >60 days of lactation could increase up to 3.1± 0.1 g/L, which was 2.18 times greater than in secretor mothers milks of same lactation period. These findings might suggest that lactose becomes more available substrate for FUT3 fucosylation due to no competition from FUT2 in non-secretor mothers.