Abstract
We present an electrochemical impedimetric-based biosensor for monitoring the variation in human milk oligosaccharide (HMO) composition. 2′-Fucosyllactose (2’FL) is an HMO associated with infant growth, cognitive development, and protection from infectious diarrhea, one of the major causes of infant death worldwide. Due to genetic variation, the milk of some women (non-secretors) contains no or very little 2′FL with potential implications for infant health and development. However, there is currently no technology to analyze the presence and concentration of HMOs in human milk at the point-of-care (POC). The lack of such technology represents a major impediment to advancing human milk research and improving maternal-infant health. Towards this unmet need, we report an impedimetric assay for HMOs with an α-1,2 linkage, the most abundant of which is 2′FL. The sensor uses a lectin for affinity, specifically Ulex europaeus agglutinin I (UEA), with electrochemical readout. In spiked studies, the sensor exhibited a high degree of linearity (R2 = 0.991) over 0.5 to 3.0 μM with a 330-nM detection limit. The sensor performance was clinically validated using banked human milk samples and correctly identified all secretor vs. non-secretor samples. Furthermore, despite the short 35-min assay time and low sample volume (25 μL), the assay was highly correlated with HPLC measurements. This bedside human milk testing assay enables POC, “sample-to-answer” quantitative HMO measurement, and will be a valuable tool to assess milk composition.
Graphical abstract
Similar content being viewed by others
Introduction
Human milk is widely accepted as the best source of nutrients for healthy term newborns [1, 2], and the composition is strongly correlated with maternal-infant health and developmental outcomes [3, 4]. Studies have found that through universal breastfeeding, the death of 823,000 children and 20,000 mothers annually could be averted along with an economic savings of $300 billion [5]. While the benefits of human milk and breastfeeding are numerous, human milk composition varies between women and is dynamic over the course of lactation. Human milk oligosaccharides (HMOs) are a group of ~150 complex glycans (sugars) and, at 5–10 g/L, represent the third most abundant solid component of human milk after lactose and lipids [6,7,8,9]. Several HMOs, like 2′-fucosyllactose (2′FL), are present in high concentrations in most women (secretors) but are nearly absent in the milk of other women (non-secretors) [10]. Differences in maternal HMO composition have been linked to several short- and long-term infant health and disease outcomes. For example, the near absence of 2′FL in the milk of non-secretor women is associated with a higher infant risk of infectious diarrhea [3, 4], which remains one of the leading causes of death in infants and children under the age of 5 worldwide [11, 12]. Higher 2′FL concentration has also been associated with improved infant growth [13, 14], improved memory and learning in rodent models, and improved infant cognitive development at 24 months in human cohorts [15, 16]. While these and other studies highlight the importance of HMOs in human milk, no point-of-care (POC) technology is currently available to quantify the amount of 2′FL in human milk.
HMOs are unconjugated sugars (i.e., they are not covalently bound to proteins or lipids) but instead carry lactose at the reducing end. Lactose can be elongated by adding one or more disaccharides and modified by adding fucose and/or sialic acid with different linkages [9]. 2′FL, lactose fucosylated with an α-1, 2 linkage at galactose, is the most abundant HMO in the milk of secretors [16]. Lectins are carbohydrate-binding proteins that recognize specific glycan structures [17, 18] and thus can be used as affinity reagents. They have been used as recognition molecules to capture glycan targets (i.e., lectin-affinity chromatography, lectin-based biosensors, etc.) [19]. Specifically, Ulex europaeus agglutinin I (UEA), which has a binding site for α-1,2 linkages [20], is a natural binding partner for 2′FL (Kd = 327 nM) [21]. We hypothesized that UEA could be used as an affinity molecule to selectively capture 2′FL out of the higher abundance HMOs present in human milk.
The lack of technology to analyze human milk bioactives at the POC represents a major impediment to advancing human milk research and improving maternal-infant health. Over the past few decades, analytical methods to detect and quantify 2′FL have been developed. Early research focused on separating 2′FL and other HMOs using chromatographic techniques [22,23,24]. Unfortunately, chromatography is still a costly and time-consuming process requiring large amounts of organic solvent and sample [25]. To overcome these disadvantages, high-performance liquid chromatography (HPLC) was coupled with other techniques (e.g., mass spectrometry (MS) [26], nuclear magnetic resonance (NMR) [27], and electrochemical techniques [28]) to detect HMOs. These approaches enhanced the sensitivity and selectivity, which allowed for the characterization of fucosylated oligosaccharides; however, they are limited to laboratory-based tests and not amenable to POC testing [29]. Currently, a significant obstacle in developing a POC testing device is that HMO analytical detection methods require burdensome sample pre-treatment, expensive instrumentation, and trained professionals to run the assays. Some macronutrients like total protein, total lipids, and lactose can be measured by “milk analyzers” at the bedside; however, the analysis of many other essential human milk components currently require specialized, expensive, and time-consuming technology that is available in only a handful of research laboratories around the world. Thus, despite the importance of HMOs to infant health, there is currently no sensor platform capable of detecting 2′FL without sending human milk samples to one of the very few research laboratories.
This study reports a rapid, label-free human milk screening assay to detect and quantify 2′FL as a biomarker to differentiate secretor vs. non-secretor status, as shown in Fig. 1. We used a lectin (UEA) covalently bound to an amine-functionalized gold electrode as the capture molecule. Upon binding with the cognate glycan (2′FL), the electrode impedance is perturbed and read out using electrochemical impedance spectroscopy (EIS). This electrochemical approach overcomes the disadvantages of the previously reported milk analysis methods. The assay requires minimal sample volume (25 μL), enabling POC, “sample-to-answer” quantitative measurement of HMOs. In summary, this assay allows a handheld POC solution for mothers and human milk studies with simple, rapid, and accurate HMO quantification.
Materials and methods
Reagents and instruments
2′-Fucosyllactose (2′FL; #SMB00933), phosphate-buffered saline (PBS; #P5493), glutaraldehyde (#G5882), sulfuric acid (#339741), cystamine dihydrochloride (#30050 sigma), human serum albumin (HSA; #A9511-100 mg), and ammonium hydroxide (#09859) were purchased from Sigma-Aldrich (USA). Disialyllacto-N-tetraose (DSLNT; #OD00172) and Ulex Europaeus agglutinin I (UEA; #L-1060) were purchased from Carbosynth (UK) and Vector Laboratories (USA), respectively. Bovine serum albumin (BSA; #37525), glucose (#A16828), Koptec pure ethanol (200 proof), and hydrogen peroxide (#HX0635-3) were obtained from Thermo Fisher Scientific (USA), Alfa Aesar (USA), Decon Labs (USA), and EMD Millipore, respectively. Bio-banked human milk samples were provided from the Mother-Milk-Infant Center of Research Excellence (MOMI CORE) at UCSD.
Voltammetry and electrochemical impedance spectroscopy measurements were performed using a benchtop potentiostat (CH Instruments, 750E), a portable multi-potentiostat (PalmSens, Palmsens4), and a custom-built, handheld potentiostat. A 3-electrode setup was used for characterization experiments with Au, Ag/AgCl, and Pt wire (CH Instruments, #CHI115) as the working (WE), reference (RE), and counter electrodes (CE), respectively. The 5-mm-diameter WE was formed with a custom fixture over a gold-sputtered glass slide (100 nm thick 99.999% pure Au film on a 5-nm Cr adhesion layer).
Biosensor fabrication
Fabrication of UEA-based sensors is a stepwise process. First, Au-coated slides were chemically cleaned using piranha solution (a 3:1 mixture of H2SO4 and H2O2), then thoroughly washed with distilled water [30]. Next, the WE was electrochemically cleaned by sweeping the potential from 0 to 1.4 V (vs. Ag/AgCl) in 0.5 M H2SO4 [31]. After washing with distilled water, the Au surface was dried with compressed air. The surface was amine-functionalized by incubating the Au electrodes with 50 μL of 250 mM cystamine dihydrochloride (Cys) solution in a 1:4 mixture of NH4OH and ethanol overnight at room temperature. After washing with ethanol, 50 μL of 25% glutaraldehyde (amine-reactive bifunctional cross-linking reagent) was drop-cast on the electrodes and incubated for 1 h. Electrodes were then washed with distilled water 3×. Immediately afterward, 50 μL of 100 μg/mL UEA was incubated with the substrate for 2 h at room temperature and 8 h at 4 °C. Lastly, the electrodes were washed with PBS and incubated with 1% BSA for 30 min. After washing the electrodes, the electrode responses were compared before quantifying 2′FL to select those having a similar response.
Sensor characterization
The stepwise modification of the electrodes was characterized using cyclic voltammetry (CV) and EIS in 0.1 M PBS (pH 7.4) containing 5 mM [Fe(CN)6]4−/3-. Voltammograms were recorded from −0.5 to 0.8 V at a 50 mV/s scan rate. The bare gold (Au), cystamine modified layer (Au/Cys), UEA-modified electrodes (Au/Cys/UEA), and BSA-treated probe (Au/Cys/UEA/BSA) were used for sensor characterization. EIS measurements were collected using a 2-electrode setup (RE and CE tied together) with an open-circuit voltage applied from 100 kHz to 0.1 Hz and an ac amplitude of 5 mV. Data from Nyquist plots were fitted with ZSimpWin software. A Randles equivalent circuit was used to extract the fitted charge transfer resistance, Rct.
Human milk assay
The 2′FL sensors were tested against a range of 2′FL concentrations by serially diluting purified 2′FL samples. The sensor was incubated with various concentrations (500 to 4500 nM) in 0.1 M PBS (pH 7.4) for 30 min at room temperature and measured using EIS with 5 mM [Fe(CN)6]4−/3- in 0.1 M PBS. Experiments were replicated (n ≥ 3) on independent sensor surfaces. As shown in Fig. 1, for the clinical samples, human milk was diluted, and 25 μL was drop-cast on the working electrode. After a 30-min incubation, EIS was performed in the same manner described previously. The change in Rct (∆Rct) was read out and used to quantify the amount of 2′FL in the samples.
Milk collection and storage
Milk samples stem from a previously reported study described by Plows et al. (2021) analyzed for HMO concentrations by HPLC, including 2’FL [32]. In brief, human milk was collected at least 1.5 h after the previous feeding and after the mother had fasted for at least 1 h. Participants provided a single full breast expression from the right breast in our clinical unit using an electric breast pump, ensuring the collection of fore-, mid-, and hind-milk. Human milk was mixed, aliquoted (into 10× 500-μL tubes, and the remainder into 5-mL tubes), and stored at −80 °C. 2′FL results quantified by HPLC were blinded to the researchers performing sensor studies and unblinded after sensor results became available. Institutional review boards at the University of Southern California and Children’s Hospital Los Angeles approved the study, where the human milk samples were collected as part of a larger study.
Statistical analysis
All data was from a minimum of three independent experiments, and error bars represent one standard deviation. Statistical analysis was performed with Origin 9.0. The limit of detection (LOD) was calculated using LOD = 3.3 × standard deviation (SD) of the regression line/slope (95% confidence level) [33]. Linear regression with Deming’s method was used to interpret the comparison data [34].
Results and discussion
Sensor surface characterization
The sensors were constructed by attaching UEA to the surface with amine chemistry. Voltammograms and spectra were recorded using CV and EIS, respectively, to characterize the stepwise sensor fabrication. CV is a valuable technique to determine the conductivity of the immobilized materials on the electrode surface using the current response of the mediator [35, 36]. Figure 2A shows voltammograms recorded at each step in the process, with the following: (a) bare gold electrode (Au), (b) addition of cystamine (Au/Cys) to present an amine functional group, (c) addition of lectin affinity reagent (Au/Cys/UEA), (d) blocking the surface with BSA (Au/Cys/UEA/BSA), and (e) the detection of the target HMO (Au/Cys/UEA/BSA/2′FL). Well-defined reversible redox peaks of ferri/ferrocyanide for the bare Au layer were observed at 128 and 284 mV. The peak current was reduced with the subsequent addition of cystamine (Au/Cys), inhibiting electron transfer, which indicated that the electrode conductivity was decreased by the covalent bond formed between Cys and Au (Au-S bonding). UEA and BSA immobilization further reduced the current response due to the accumulation of non-conductive proteins on the surface. Lastly, adding 250 nM of 2′FL to the sensor slightly decreased the peak current, indicating binding.
Next, EIS spectra were recorded to study the surface properties of the sensor, as quantified by the charge transfer resistance, Rct, since this is the most sensitive parameter to characterize what is happening on the electrode surface [35, 37]. To obtain the Rct values, the Nyquist plot was fitted with a Randles equivalent circuit model (inset of Fig. 2B). As shown in Fig. 2B, Rct increased with the stepwise immobilization due to the adsorbed non-conductive substances on the sensor surface. Because of the high electron transfer rate, the bare Au sensor had the lowest Rct (31.4 ± 2.2 Ω). After adding the self-assembling Cys, Rct increased to 118.3 ± 6.5 Ω due to increased steric hindrance and further increased when UEA was immobilized (519.4 ± 13.8 Ω), and the surface was blocked with BSA (558.8 ± 11.4 Ω) to mitigate non-specific binding. These data indicate that the capture probe was successfully attached to the sensor surface. These values are consistent with the CV responses. Finally, Fig. 2C shows contact angle measurements made at different stages of the modification process. The piranha-cleaned Au surfaces are hydrophilic (Fig. 2C, i), where acid treatments are widely used to eliminate organic contaminants and generate a hydrophilic surface. The self-assembled Cys monolayer (Fig. 2C, ii) had an increased contact angle (44°) because of the successful monolayer formation. Figure 2C, iii, shows the contact angle (76°) after adding UEA, indicating that the surface is now relatively hydrophobic. These results collectively indicate that the sensor is well-formed and demonstrates that it is responsive to the target HMO.
Optimization
To maximize the response towards 2′FL, the assay parameters were systematically optimized. Specifically, the UEA immobilization time, 2′FL incubation time, and assay pH were studied. EIS was used to assess the effect of each parameter by quantifying the change in Rct (ΔRct). As shown in Electronic Supplementary Material Fig. S1A, EIS was performed at various times during UEA immobilization on the Au/Cys layer. For the UEA incubation time, ΔRct monotonically increased from 3 to 10 h and then plateaued. Thus, 10 h was chosen as the optimal condition. The target molecule incubation time, the critical factor in deciding the assay time, was investigated from 5 to 45 min (see Electronic Supplementary Material Fig. S1B) similarly. ΔRct increased linearly from 5 to 20 min and leveled off after 30 min, likely due to saturating all the available binding sites. Finally, the effect of measurement buffer pH was studied, ranging from 5.5 to 8.0, where pH 7.4 showed the maximum response (see Electronic Supplementary Material Fig. S1C). These optimum assay conditions were used for all subsequent experiments.
Interference
Human milk is a complex sample with an abundance of proteins, lipids, and sugars. The target HMO, 2′FL, is present at 2–5 g/L, whereas there is 5–15 g/L of total HMO [8], along with proteins at 8 g/L, fat at 41 g/L, 0.25 mg/mL of glucose [38], etc., as shown in Fig. 3A. HMOs are diverse with over 60% being fucosylated (e.g., 2′FL, 3-fucosyllactose (3FL), lacto-N-fucopentaose I (LNFP I)), 13% sialylated (e.g., 3′-sialyllactose (3′SL), 6′-sialyllactose (6′SL)), and some without any modifications (e.g., lacto-N-tetraose (LNT), lacto-N-neotetraose (LNnT)), etc. Notably, non-secretors have significantly lower or absent 2′FL and LNFP 1 due to the inactive FUT2 enzyme [39].
To study the effect of these off-target molecules’ impact on the assay performance, we measured the sensor response in the presence of glucose (0.5 mg/mL; 2.78 mM), 10× diluted bovine milk (3.3 g/L of proteins, 0.1–0.2 g/L of bovine milk oligosaccharide [40, 41]), HSA (1.0 mg/mL; 15.1 μM), disialyllacto-N-tetraose (DSLNT; 1.0 μM), and 2′FL (1.0 μM). As shown in Fig. 3B, the sensor responses were negligible to these interfering species (P value <0.05), and the sensor showed a strong response to 2′FL. This experiment has two essential negative controls: bovine milk, which contains HMO analogs [42], and DSLNT, another HMO that, unlike 2′FL, is not α-1-2-fucosylated, but carries sialic acid instead. Critically, the sensor showed minimal response to these off-target glycans and robust signal to 2′FL. To further validate this and prove that the interaction is specific, we measured the sensor response to 2′FL and DLSNT at 1.0 μM on sensors with and without the affinity lectin, UEA. As shown in Fig. 3C, in the absence of UEA (red bars), the sensor showed a similar response to both 2′FL (12.3 ± 1.6 Ω) and DSLNT (11.0 ± 5.8 Ω). Meanwhile, with the affinity lectin immobilized, the sensor exhibited a strong signal for 2′FL (177.1 ± 3.9 Ω) and very little signal for DSLNT (8.5 ± 4.6 Ω). These results indicate that the sensor can selectively detect 2′FL by recognizing the ɑ-1,2 linkage of 2′FL and is not sensitive to off-target molecules.
Detection of 2′FL
The sensor’s analytical performance was investigated using EIS in 0.1 M PBS (pH 7.4) containing ferri-/ferro-cyanide. For each measurement, a stock sample containing 2′FL in PBS was serially diluted and assayed by placing 25 μL on the sensor and incubating for 30 min. As shown in Fig. 4A and B, the Nyquist plots showed a linear proportionality between Rct and the 2′FL concentration from 0.5 to 3.0 μM. The linear regression equation for the plot in the standard sample is ΔRct = 60.438(±7.035) + 0.114(±0.005) (2′FL in nM) with a correlation coefficient of 0.991. The limit of detection (LOD) is 330.7 nM (161.5 ng/mL), which is superior to the previously reported methods [23, 28, 43, 44]. To define a secretor and non-secretor, a cutoff of 410 μM was chosen based on previous reports [45]. Therefore, the developed 2′FL sensor can be used as a milk analyzer to distinguish secretor/non-secretor status using 1000-fold diluted milk samples. As shown in Fig. 4C and D, the developed sensor was evaluated with 2′FL spiked (0.5, 1.0, and 2.0 μM) in human milk samples (non-secretor) diluted 1000-fold. Despite the more complex milk matrix, the measured Rct values were still proportional to the 2′FL concentration. The linear regression equation is ΔRct = 45.323(±8.035) + 0.120(±0.01) (2′FL in nM) with a correlation coefficient of 0.98. Compared to the PBS samples, the milk response was slightly increased, which is likely due to the complexity of the milk matrix and the existence of unknown interfering species adsorbing on the sensor.
Reproducibility and long-term storage stability
Sensor fabrication consistency is essential to achieve reproducibility [46], which can be achieved by improving the uniformity of the self-assembled monolayers (SAM) and the covalently attached receptors on the sensor surface. In particular, the sensor surface cleanness and reagent contamination significantly affect the SAM uniformity and cross-linked receptors [47]. Therefore, the substrates were thoroughly washed with chemical and electrochemical methods prior to sensor fabrication using freshly prepared reagents. EIS was performed on five independent sensors produced from the same batch of substrates, and the coefficient of variation (CV) was calculated to assess the consistency. Figure 5A shows the measured responses to 1.0 μM of 2′FL where the CV was less than 4.7% for 15 measurements, indicating the sensor fabrication and response were highly reproducible. Next, the sensor’s long-term storage stability was evaluated by measuring the sensor response over a 2-week period. Sensors were prepared and then stored in a refrigerator at 4 °C and covered with parafilm between measurements. The sensors were periodically removed, incubated with 1.0 μM of 2′FL, measured, and washed. As shown in Fig. 5B, the sensors retained more than 95% of their initial day 0 response for up to 10 days, 92% at 12 days, and decreased slightly to 88% after 14 days. These results demonstrate that the Au/Cys/UEA/BSA functionalized sensor has high reproducibility and maintains stability over a long period. For binary detection (secretor vs. non-secretor), the sensors are likely stable for an even longer time, although that was not explored in this work.
Clinical validation
To verify the clinical applicability, donor human milk samples were quantified using the developed electrochemical sensors. Banked human milk samples were obtained from the Mother-Milk-Infant Center of Research Excellence (MOMI CORE) at the University of California San Diego and prepared as described above. Samples (n = 12) were provided blinded and tested using the proposed assay. As shown in Fig. 6A, 4 of the 12 samples (A, F, I, and K) had signals below the cutoff threshold and were identified as non-secretor milk, while the others were called secretors. Those identified as secretors (B, C, D, E, G, H, J, and L) were then 10× further diluted and re-run to bring their signal within the sensor’s dynamic range for quantification. After unblinding the samples, all agreed with the HPLC data (100% positive percent agreement and 100% negative percent agreement). While the sample size is relatively small, the concentration range is representative and highly encouraging. Furthermore, the error bars on the replicate samples are small, indicating little variability in the assay. The EIS results were compared against the gold standard measurement technique (HPLC) run on the same samples (Electronic Supplementary Material Table S1). The agreement between the methods was evaluated using a Deming method comparison test [34]. As shown in Fig. 6B, there was no significant difference between the two analytical methods as the EIS results exhibit a strong correlation with the HPLC data (R2 = 0.97). The data were also evaluated using the paired t-test, and the calculated t value (1.42) was less than the critical t value (2.365) at the 95% confidence level (n = 8). It is worth pointing out that HPLC is selective against just 2′FL, whereas this test measures all HMOs with an α-1,2 linkage, of which 2′FL is the most abundant. These data demonstrate that the proposed sensor can accurately quantify 2′FL and be used to differentiate secretor vs. non-secretors at the bedside.
Comparison
Over the last several decades, various analytical methods to quantify 2′FL have been developed, such as gel filtration with paper chromatography, HPLC, hydrophilic interaction liquid chromatography with fluorescence detection (HILIC-FLD), high-performance anion-exchange chromatography (HPAEC), and a whole-cell biosensor with fluorescence detection. Their performance is summarized and compared in Table 1. The developed impedimetric 2′FL sensor has the lowest limit of detection (LOD) and the fastest assay time (35 min). Notably, the proposed sensor requires no preprocessing steps (such as centrifugation), enabling it to be used for point-of-care testing. In addition, this work is the most cost-effective (<$3/test) (see Electronic Supplementary Material Table S2 for a cost breakdown). While electrochemical techniques have an affordable instrumental cost considering the device portability (<$5000), HPLC and multiplexed HPLC methods require a specific detector and column for each purpose and high instrument cost (~$27,500) [48]. The reader for the developed assay can be miniaturized [49,50,51], enabling screening human milk on-site and hence used as a bedside milk analyzer.
Conclusion
This work demonstrated a label-free, rapid, simple, low-cost analysis platform to differentiate secretor and non-secretor status in human milk. Selective detection is accomplished using a lectin affinity reagent towards HMOs with α1-2-fucosylation, such as the highly abundant 2′-fucosyllactose. The sensor is read out using an electrochemical impedimetric assay where 2′FL selectively binds to UEA and decreases the charge transfer resistance. This scheme was optimized and then validated using PBS and non-secretor donor milk spiked with 2′FL. The assay specificity against off-target HMOs and potential interferers was verified, with little to no cross-reactivity observed. The reproducibility and stability were demonstrated over a 2-week timeframe with a low coefficient of variation. This set the stage for clinical validation where donor human milk was screened for secretor status. The results were in perfect agreement with HPLC despite the ~4× shorter assay time and significantly lower cost. A fast HMO assay with a low sample volume (<25 μL) enables on-site, bedside human milk monitoring throughout lactation and could result in better infant health outcomes. This test could also be used for donor human milk screening to decide on the use of secretor vs. non-secretor milk once research studies confirm the benefits of one over the other depending on a given situation and context. While the proposed assay offers considerable promise to manage variation in human milk composition due to its simplicity, user-friendly, and fast response time, it requires further clinical validation with a larger sample size. Future work will expand this strategy to other HMOs and human milk bioactives for a more holistic milk composition assessment. This assay provides a new perspective in milk research and lays the foundation for a POC bedside milk analyzer for clinical purposes and a milk screening tool.
References
Ballard O, Morrow AL. Human milk composition: nutrients and bioactive factors. Pediatr Clin N Am. 2013;60:49–74. https://doi.org/10.1016/j.pcl.2012.10.002.
Underwood MA. Human milk for the premature infant. Pediatr Clin N Am. 2013;60:189–207. https://doi.org/10.1016/j.pcl.2012.09.008.
Ruiz-Palacios GM, Cervantes LE, Ramos P, Chavez-Munguia B, Newburg DS. Campylobacter jejuni binds intestinal H(O) antigen (Fucα1, 2Galβ1, 4GlcNAc), and fucosyloligosaccharides of human milk inhibit its binding and infection. J Biol Chem. 2003;278:14112–20. https://doi.org/10.1074/jbc.M207744200.
Morrow AL, Ruiz-Palacios GM, Altaye M, Jiang X, Guerrero ML, Meinzen-Derr JK, Farkas T, Chaturvedi P, Pickering LK, Newburg DS. Human milk oligosaccharides are associated with protection against diarrhea in breast-fed infants. J Pediatr. 2004;145:297–303. https://doi.org/10.1016/j.jpeds.2004.04.054.
(2016) The Lancet: Increasing breastfeeding worldwide could prevent over 800,000 child deaths every year. In: Baby Friendly Initiative. https://www.unicef.org.uk/babyfriendly/lancet-increasing-breastfeeding-worldwide-prevent-800000-child-deaths-every-year/. Accessed 13 Jun 2020.
Newburg DS, Ruiz-Palacios GM, Morrow AL. Human milk glycans protect infants against enteric pathogens. Annu Rev Nutr. 2005;25:37–58. https://doi.org/10.1146/annurev.nutr.25.050304.092553.
Bode L. Recent advances on structure, metabolism, and function of human milk oligosaccharides. J Nutr. 2006;136:2127–30. https://doi.org/10.1093/jn/136.8.2127.
Bode L. Human milk oligosaccharides: every baby needs a sugar mama. Glycobiology. 2012;22:1147–62. https://doi.org/10.1093/glycob/cws074.
Bode L, Jantscher-Krenn E. Structure-function relationships of human milk oligosaccharides123. Adv Nutr. 2012;3:383S–91S. https://doi.org/10.3945/an.111.001404.
Smith-Brown P, Morrison M, Krause L, Davies PSW. Mothers secretor status affects development of childrens microbiota composition and function: a pilot study. PLoS One. 2016;11. https://doi.org/10.1371/journal.pone.0161211.
Vandenplas Y, Berger B, Carnielli VP, Ksiazyk J, Lagström H, Sanchez Luna M, Migacheva N, Mosselmans J-M, Picaud J-C, Possner M, Singhal A, Wabitsch M. Human milk oligosaccharides: 2′-fucosyllactose (2′-FL) and lacto-N-neotetraose(LNnT) in infant formula. Nutrients. 2018;10:1161. https://doi.org/10.3390/nu10091161.
Ramani S, Stewart CJ, Laucirica DR, Ajami NJ, Robertson B, Autran CA, Shinge D, Rani S, Anandan S, Hu L, Ferreon JC, Kuruvilla KA, Petrosino JF, Venkataram Prasad BV, Bode L, Kang G, Estes MK. Human milk oligosaccharides, milk microbiome and infant gut microbiome modulate neonatal rotavirus infection. Nat Commun. 2018;9. https://doi.org/10.1038/s41467-018-07476-4.
Larsson MW, Lind MV, Laursen RP, Yonemitsu C, Larnkjær A, Mølgaard C, Michaelsen KF, Bode L (2019) Human milk oligosaccharide composition is associated with excessive weight gain during exclusive breastfeeding—an explorative study. Front Pediatr 7: . https://doi.org/10.3389/fped.2019.00297.
Alderete TL, Autran C, Brekke BE, Knight R, Bode L, Goran MI, Fields DA. Associations between human milk oligosaccharides and infant body composition in the first 6 mo of life. Am J Clin Nutr. 2015;102:1381–8. https://doi.org/10.3945/ajcn.115.115451.
Berger PK, Plows JF, Jones RB, Alderete TL, Yonemitsu C, Poulsen M, Ryoo JH, Peterson BS, Bode L, Goran MI. Human milk oligosaccharide 2′-fucosyllactose links feedings at 1 month to cognitive development at 24 months in infants of normal and overweight mothers. PLoS One. 2020;15:e0228323. https://doi.org/10.1371/journal.pone.0228323.
Oliveros E, Martín MJ, Torres-Espínola FJ, Segura-Moreno MT, Ramírez M, Santos A, Rueda R, Escudero M, Catena A, Azaryah H, Campoy C (2021) Human milk levels of 2´-fucosyllactose and 6´-sialyllactose are positively associated with infant neurodevelopment and are not impacted by maternal BMI or diabetic status. 4:11.
Lis H, Sharon N. Lectin-carbohydrate interactions. Curr Opin Struct Biol. 1991;1:741–9. https://doi.org/10.1016/0959-440X(91)90173-Q.
Hirabayashi J. Concept, strategy and realization of lectin-based glycan profiling. J Biochem. 2008;144:139–47. https://doi.org/10.1093/jb/mvn043.
Silva MLS (2019) Chapter one - lectin biosensors in cancer glycan biomarker detection. In: Makowski GS (ed) Advances in clinical chemistry. Elsevier, pp 1–61.
Kondoh D, Kamikawa A, Sasaki M, Kitamura N. Localization of α1-2 fucose glycan in the mouse olfactory pathway. CTO. 2017;203:20–8. https://doi.org/10.1159/000447009.
Li X, Song S, Shuai Q, Pei Y, Aastrup T, Pei Y, Pei Z. Real-time and label-free analysis of binding thermodynamics of carbohydrate-protein interactions on unfixed cancer cell surfaces using a QCM biosensor. Sci Rep. 2015;5:14066. https://doi.org/10.1038/srep14066.
Schmid D, Behnke B, Metzger J, Kuhn R. Nano-HPLC–mass spectrometry and MEKC for the analysis of oligosaccharides from human milk. Biomed Chromatogr. 2002;16:151–6. https://doi.org/10.1002/bmc.152.
Christensen AS, Skov SH, Lendal SE, Hornshøj BH. Quantifying the human milk oligosaccharides 2′-fucosyllactose and 3-fucosyllactose in different food applications by high-performance liquid chromatography with refractive index detection. J Food Sci. 2020;85:332–9. https://doi.org/10.1111/1750-3841.15005.
Ruhaak LR, Lebrilla CB. Advances in analysis of human milk oligosaccharides123. Adv Nutr. 2012;3:406S–14S. https://doi.org/10.3945/an.112.001883.
(2006) Advantages and disadvantages of HPLC. In: HPLC. John Wiley & Sons, Ltd, pp 1–13.
Tonon KM, Miranda A, Abrão ACFV, de Morais MB, Morais TB. Validation and application of a method for the simultaneous absolute quantification of 16 neutral and acidic human milk oligosaccharides by graphitized carbon liquid chromatography – electrospray ionization – mass spectrometry. Food Chem. 2019;274:691–7. https://doi.org/10.1016/j.foodchem.2018.09.036.
van Leeuwen SS, Schoemaker RJW, Gerwig GJ, van Leusen-van Kan EJM, Dijkhuizen L, Kamerling JP. Rapid milk group classification by 1H NMR analysis of Le and H epitopes in human milk oligosaccharide donor samples. Glycobiology. 2014;24:728–39. https://doi.org/10.1093/glycob/cwu036.
Austin S, Cuany D, Michaud J, Diehl B, Casado B. Determination of 2′-fucosyllactose and lacto-N-neotetraose in infant formula. Molecules. 2018;23. https://doi.org/10.3390/molecules23102650.
Manikandan VS, Adhikari B, Chen A. Nanomaterial based electrochemical sensors for the safety and quality control of food and beverages. Analyst. 2018;143:4537–54. https://doi.org/10.1039/C8AN00497H.
Yun Y, Pan M, Wang L, Li S, Wang Y, Gu Y, Yang J, Wang S. Fabrication and evaluation of a label-free piezoelectric immunosensor for sensitive and selective detection of amantadine in foods of animal origin. Anal Bioanal Chem. 2019;411:5745–53. https://doi.org/10.1007/s00216-019-01954-4.
Miao P, Shen M, Ning L, Chen G, Yin Y. Functionalization of platinum nanoparticles for electrochemical detection of nitrite. Anal Bioanal Chem. 2011;399:2407–11. https://doi.org/10.1007/s00216-010-4642-3.
Plows JF, Berger PK, Jones RB, Alderete TL, Yonemitsu C, Najera JA, Khwajazada S, Bode L, Goran MI. Longitudinal changes in human milk oligosaccharides (HMOs) over the course of 24 months of lactation. J Nutr. 2021;151:876–82. https://doi.org/10.1093/jn/nxaa427.
Grawe GF, de Oliveira TR, de Andrade NE, Moccelini SK, Terezo AJ, Soares MA, Castilho M. Electrochemical biosensor for carbofuran pesticide based on esterases from Eupenicillium shearii FREI-39 endophytic fungus. Biosens Bioelectron. 2015;63:407–13. https://doi.org/10.1016/j.bios.2014.07.069.
Chen Z, Fang C, Wang H, He J, Deng Z. A novel disposable amperometric UA strip. Sensors Actuators B Chem. 2008;129:710–5. https://doi.org/10.1016/j.snb.2007.09.061.
Li Y, Xie G, Qiu J, Zhou D, Gou D, Tao Y, Li Y, Chen H. A new biosensor based on the recognition of phages and the signal amplification of organic-inorganic hybrid nanoflowers for discriminating and quantitating live pathogenic bacteria in urine. Sensors Actuators B Chem. 2018;258:803–12. https://doi.org/10.1016/j.snb.2017.11.155.
Etienne M, Guillemin Y, Grosso D, Walcarius A. Electrochemical approaches for the fabrication and/or characterization of pure and hybrid templated mesoporous oxide thin films: a review. Anal Bioanal Chem. 2013;405:1497–512. https://doi.org/10.1007/s00216-012-6334-7.
Serrano PC, Nunes GE, Avila LB, Reis CPS, Gomes AMC, Reis FT, Sartorelli ML, Melegari SP, Matias WG, Bechtold IH. Electrochemical impedance biosensor for detection of saxitoxin in aqueous solution. Anal Bioanal Chem. 2021. https://doi.org/10.1007/s00216-021-03603-1.
Goran MI, Martin AA, Alderete TL, Fujiwara H, Fields DA (2017) Fructose in breast milk is positively associated with infant body composition at 6 months of age. Nutrients 9: . https://doi.org/10.3390/nu9020146.
Seferovic MD, Mohammad M, Pace RM, Engevik M, Versalovic J, Bode L, Haymond M, Aagaard KM. Maternal diet alters human milk oligosaccharide composition with implications for the milk metagenome. Sci Rep. 2020;10:22092. https://doi.org/10.1038/s41598-020-79022-6.
Foroutan A, Guo AC, Vazquez-Fresno R, Lipfert M, Zhang L, Zheng J, Badran H, Budinski Z, Mandal R, Ametaj BN, Wishart DS. Chemical composition of commercial cow’s milk. J Agric Food Chem. 2019;67:4897–914. https://doi.org/10.1021/acs.jafc.9b00204.
Raikos V. Effect of heat treatment on milk protein functionality at emulsion interfaces. A review. Food Hydrocoll. 2010;24:259–65. https://doi.org/10.1016/j.foodhyd.2009.10.014.
Zivkovic AM, Barile D. Bovine milk as a source of functional oligosaccharides for improving human health. Adv Nutr. 2011;2:284–9. https://doi.org/10.3945/an.111.000455.
Enam F, Mansell TJ. Linkage-specific detection and metabolism of human milk oligosaccharides in Escherichia coli. Cell Chemical Biology. 2018;25:1292–1303.e4. https://doi.org/10.1016/j.chembiol.2018.06.002.
Goehring KC, Kennedy AD, Prieto PA, Buck RH. Direct evidence for the presence of human milk oligosaccharides in the circulation of breastfed infants. PLoS One. 2014;9:e101692. https://doi.org/10.1371/journal.pone.0101692.
Wu J, Wu S, Huo J, Ruan H, Xu X, Hao Z, Wei Y. Systematic characterization and longitudinal study reveal distinguishing features of human milk oligosaccharides in China. Curr Dev Nutr. 2020;4. https://doi.org/10.1093/cdn/nzaa113.
Chen L-C, Wang E, Tai C-S, Chiu Y-C, Li C-W, Lin Y-R, Lee T-H, Huang C-W, Chen J-C, Chen WL. Improving the reproducibility, accuracy, and stability of an electrochemical biosensor platform for point-of-care use. Biosens Bioelectron. 2020;155:112111. https://doi.org/10.1016/j.bios.2020.112111.
Wang L, Veselinovic M, Yang L, Geiss BJ, Dandy DS, Chen T. A sensitive DNA capacitive biosensor using interdigitated electrodes. Biosens Bioelectron. 2017;87:646–53. https://doi.org/10.1016/j.bios.2016.09.006.
Soares MCP, Cabral TD, Mendes BF, da Silva VA, Tambourgi EB, Fujiwara E. Technical and economic viability analysis of optical fiber sensors for monitoring industrial bioreactors. Engineering Proceedings. 2020;2:57. https://doi.org/10.3390/ecsa-7-08161.
Sun AC, Hall DA. Point-of-caresmartphone-based electrochemical biosensing. Electroanalysis. 2019;31:2–16. https://doi.org/10.1002/elan.201800474.
Jiang H, Sun A, Venkatesh AG, Hall DA. An audio jack-based electrochemical impedance spectroscopy sensor for point-of-care diagnostics. IEEE Sensors J. 2016;17:589–97.
Sun A, Venkatesh AG, Hall DA. A multi-technique reconfigurable electrochemical biosensor: enabling personal health monitoring in mobile devices. IEEE Transactions on Biomedical Circuits and Systems. 2016;10:945–54. https://doi.org/10.1109/TBCAS.2016.2586504.
Kobata A. Possible application of milk oligosaccharides for drug development. Chang Gung Med J. 2003;26:621–36.
Acknowledgements
We thank Prof. Kamil Godula and Dr. Nissi Varki for their helpful discussions and the volunteers who provided samples.
Funding
This work was partially supported by funding from the Larsson-Rosenquist Foundation Mother-Milk-Infant Center of Research Excellence (MOMI CORE) and the Altman Clinical & Translational Research Institute (ACTRI) at the University of California, San Diego. The ACTRI is funded by the National Institutes of Health (NIH) National Center for Advancing Translational Sciences (UL1TR0001442-06).
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare no competing interests.
Additional information
Published in the topical collection Point-of-Care Testing with guest editors Oliver Hayden, Peter B. Luppa, and Junhong Min.
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
ESM 1
(DOCX 267 kb)
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Chung, S., Bode, L. & Hall, D.A. Point-of-care human milk testing for maternal secretor status. Anal Bioanal Chem 414, 3187–3196 (2022). https://doi.org/10.1007/s00216-021-03697-7
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00216-021-03697-7