An Efficient Heparin Affinity Column Purification Method Coupled with Ultraperformance Liquid Chromatography for the Quantification of Native Lactoferrin in Breast Milk

Lactoferrin (LF) is a bioactive multifunctional protein and found in the highest amounts in human milk. Several methods can be used to quantify LF. However, quantiﬁcation of native LF has garnered relatively little interest to date. This study aimed to develop a novel eﬃcient two-step method for quantifying native LF in breast milk. During the analysis, LF was ﬁrst extracted with phosphate buﬀer (pH 5.0), puriﬁed using a heparin aﬃnity column. Subsequently, LF was detected using ultraperformance liquid chromatography (UPLC) at a wavelength of 201nm. A linear calibration curve was obtained in the range of 5–200 mg/L. The limit of detection and limit of quantitation were 1mg/L and 5 mg/L, respectively, indicating that the validated method could be employed to quantify LF in breast milk. Compared with previous HPLC methods, this method demonstrated several remarkable advantages, including simple operation, low-cost detection, and high accuracy. Hence, the results demonstrate an eﬃcient method that can be employed commercially to purify and analyze LF in human milk samples.


Introduction
Lactoferrin (LF) is a multifunctional globular glycoprotein that can be characterized as a member of the transferrin family [1,2]. It is an iron-binding protein composed of a single-chain polypeptide with two globular lobes [3]. e polypeptide chain consists of approximately 600-700 amino acid residues and has a molecular weight of 80 kDa [4]. LF exhibits a variety of functional properties due to its unique structural features. e antibacterial activity is one of the most critical functions of LF [5]. LF also mediates other biological activities, such as regulation of iron transport, immune system, transcription, proteolysis, and enzyme activity, along with antitumor and anti-inflammatory effects [6]. e richest sources of LF are breast and bovine milk. However, many differences have been observed in LF present in human and bovine milk. e concentration of LF in milk varies significantly with the lactation period and species [7,8]. LF concentration in bovine milk (approximately 0.8-1.0 g/L in colostrum and 0.03-0.49 g/L in mature milk) [9] is lower than that in human milk and its colostrum (approximately 5.0-7.0 g/L in colostrum and 1.0-3.0 g/L in mature milk) [10]. Structurally, LF in human and bovine sources exhibits approximately 70% sequence homology. Furthermore, LF in human and bovine milk is composed of 691 and 696 amino acids, respectively [11]. Due to differences between bovine LF and human LF, accurate and rapid determination of LF in human milk is of great significance for nutritional research and commercial design of infant formula.
Common quantification methods can be classified into immunological and nonimmunological methods [12]. Immunological methods include enzyme-linked immunosorbent assay (ELISA) and surface plasmon resonance-based immunosensors [13]. Although immunological methods exhibit high selectivity and sensitivity, they have poor reproducibility [14]. Nonimmunological techniques such as ultraperformance liquid chromatography-tandem mass spectrometry (UPLC-MS), reversed-phase high-performance liquid chromatography (RP-HPLC), and capillary electrophoresis (CE) have also been reported [6]. UPLC-MS enables highly sensitive and selective detection of LF using multiple reaction monitoring modes based on signature peptides [12]. RP-HPLC can be used for the determination of native LF. However, a sensitive detection can only be observed when evaluating dairy products with high LF content [15,16]. CE is widely used for protein analysis; however, it is limited by capillary, voltage, and other conditions. Hence, it may not be sensitive enough to separate proteins with small differences in molecular weight [17]. Intact LF remains indistinguishable from its denatured form when analyzed using chromatographic techniques, which may lead to the overestimation of LF. Heparin, a sulfated polysaccharide belonging to the glycosaminoglycan family, is used to enrich native LF [18]. Heparin has the highest negative charge density among known biological macromolecules [19]. erefore, heparin exhibits a high binding affinity to various proteins, including LF but excluding ß-lactoglobulin, α-lactalbumin, and serum albumin [16]. Heparin affinity column has been widely used for the purification of LF due to its special properties. In the study performed by Lutaty et al. [20], the breast milk samples were loaded onto a heparin affinity column and washed and eluted with different salt concentrations. Subsequently, the protein was collected, thereby improving the purity of the LF-enriched fraction. Similarly, LF was purified from transgenic milk and bovine milk retentate using affinity chromatography as described by Parc et al. [21].
Herein, the objective of the study was to develop an efficient method using heparin affinity column purification coupled with UPLC detection for the determination of native LF in breast milk. A heparin affinity column was employed for the purification of LF by optimizing extraction conditions and elution conditions.

Reagents and Instruments.
Heparin affinity column was purchased from Meizheng Bio-Tech Co., Ltd. (Beijing, China). Disodium hydrogen phosphate dodecahydrate, sodium dihydrogen phosphate dihydrate, sodium chloride, and trifluoroacetic acid were purchased from Damao Chemical Reagent Factory (Tianjin, China). Acetonitrile and methanol were purchased from Merck (Darmstadt, Germany). Human LF (≥85%) was purchased from Sigma-Aldrich (St. Louis, MO, USA). Deionized water was purified using the Milli-Q water purification system (Millipore, MA, USA). HPLC grade reagents were used in this study.

Sample Collection.
Healthy mothers with ages ranging from 20 to 35 years participated in this study. Women with physical or mental disorders and those under nutritional or drug intervention were excluded from this study. Each participant submitted written informed consent prior to inclusion in the study. A total of three colostrum, two transitional milk, and 12 mature milk samples were collected using a manual suction pump between 09:00 a.m. and 11:00 a.m. from day 1 to day 330 after delivery. After collection, the samples were frozen and stored at −80°C until further analysis.

Dynamic Adsorption Analysis Using a Heparin Affinity
Column.
e working principle of the heparin affinity column is depicted in Figure 1(a). Continuous flow adsorption studies were conducted using an adsorption column with an inner diameter of 1.5 cm and a height of 12 cm. Before the sample injection, the column was flushed with buffer. During the process of injection, the sample was added to the top, which flowed through the column owing to the effect of gravitational force. e effluent samples were collected at specified times and analyzed using UPLC. e dynamic behavior of the columns was predicted using logistic models. e linearized model for a singlecomponent system was expressed using the following equation:

Sample Extraction.
A total of 1 mL of breast milk was centrifuged at 10000 rpm for 8 min at 4°C [6]. After centrifugation, the supernatant was transferred to a new centrifuge tube. Buffer solution A (99.38 mmol/L Na 2 HPO 4 ·12H 2 O and 0.61 mmol/L NaH 2 PO 4 ·2H 2 O, pH 5.0) was added, followed by mixing the sample using a vortex mixer for 2 min. e sample extract was loaded onto the column as a loading solution.

Sample Cleanup.
A total of 5 ml of buffer solution A was added to the heparin affinity column. e sample extract was added after treatment with the buffer. After the sample solution was eluted completely, the column was rinsed with 10 mL of buffer solution B (99.38 mmol/L Na 2 HPO 4 ·12H 2 O, 0.61 mmol/L NaH 2 PO 4 ·2H 2 O, and 100 mmol/L NaCl). All effluents were discarded. Finally, the column was eluted with 4 mL of eluent (99.38 mmol/L Na 2 HPO 4 ·12H 2 O, 0.61 mmol/L NaH 2 PO 4 ·2H 2 O, and 1 mol/L NaCl) [22]. e eluent was collected and centrifuged at 9000 rpm for 5 min before UPLC analysis.
e UPLC system was controlled with Waters® Empower TM chromatography software. e column temperature was maintained at 60°C. e UV detection wavelength and the injection amount were maintained at 201 nm and 10 μL, respectively. Line A in the mobile phase involved 0.1% trifluoride acetic acid mixed in water, whereas line B involved 0.09% trifluoroacetic acid mixed in acetonitrile solution. Gradients were set as shown in Table 1.

Statistical Analysis.
All experiments were performed in triplicate, and the results are expressed as mean ± standard deviation. Origin 9.0 (OriginLab Corporation, USA) was used to obtain scientific graphs, and SPSS Statistics 22.0 was used to determine significant differences between samples with p < 0.05.    Figure 1(b) shows the breakthrough curve determined during the continuous adsorption experiment performed using the heparin affinity column. e breakthrough curve showed an exceptionally good shape, which indicated that LF finally reached adsorption equilibrium with effective adsorption on the adsorption column.

Optimization of Chromatographic Conditions.
e choice of wavelength can improve the detection of human LF (HLF) by maximizing the sensitivity and minimizing interferences. As shown in Figures 2(a) and 2(b), HLF was detected at a wavelength ranging from 200.5 to 201.7 nm. According to the peak responsivity, 201 nm was selected as the detection wavelength.
Some studies have demonstrated a high UV response of HLF at 280 nm [23,24]. Hence, the wavelength was determined by comparing the UV absorbance measured at 201 nm and 280 nm. e UV absorbances measured at 201 and 280 nm were monitored for the determination of the protein elution profile. As shown in Figure 2(c) and Supplementary Table 1, the relationship between the relative protein concentration in human milk and the signal observed at 280 nm was linear. However, many peaks corresponding to impurities were observed in the sample eluent. During the detection of HLF at 201 nm, the response sensitivity was significantly improved due to a better resolution of the HLF peak at the baseline. erefore, 201 nm was selected as the detection wavelength. e choice of the chromatography column can influence the separation of HLF from sample components [25]. Here, a C4 column was selected for the HLF detection.

Effect of the Ratio of the Sample to Buffer Solution A.
An appropriate extraction rate is conducive to the extraction of HLF [26]. Considering the extraction efficiency, four dilutions of 1 : 20, 1 : 50, 1 : 100, and 1 : 200 were used for analysis, while other variables were set as follows: pH of buffer solution A was maintained at 7.5. e phosphate concentration in buffer solution A was 100 mmol/L. NaCl concentration in buffer solution B was 100 mmol/L. NaCl concentration in the eluent was 1.0 mol/L. e influence of sample dilution on the measurement of HLF was analyzed.
e effect of dilution can be observed in Figure 3(c). At the extraction rate observed with a dilution of 1 : 100, the value was approximately 88%.

Effect of Phosphate Concentrations in Buffer Solution A.
e target protein in the sample was extracted using buffer solution A. Retaining the activity of the protein throughout the process was crucial. To study the effect of buffer solution A on HLF activity and extraction efficiency, different concentrations of phosphate (50, 100, and 200 mmol/L) were tested during the measurement. e results showed that the optimal phosphate concentration was 100 mmol/L. A significant decrease was observed in the LF reaction at all concentrations higher than the optimal phosphate concentration (Figure 3(b)). is might be attributed to the fact that the binding of LF to heparin involved an electrostatic interaction, which could be disrupted with an increase in the ionic strength [16].

Effect of pH Values in Buffer Solution
A. pH of buffer solution A has a significant effect on the structure and stability of LF [27]. In this study, aqueous phase pH was varied between 3.0 and 9.0 to study the effect of pH on the HLF extraction. e extraction efficiency was very low at basic pH, but it was found to increase in acidic pH (Figures 3(c) and 3(f )). erefore, pH 5.0 was selected to further optimize purification conditions. e thermal stability of LF was severely affected, and it gelled at neutral and alkaline pH when heated at 80-100°C for 5 min [28]. Under acidic conditions, LF was relatively stable when heated at approximately pH 3.0-4.0 [28]. is implies that LF exhibits better thermal stability under acidic conditions.

Effect of NaCl Concentrations in Buffer Solution B.
e purpose of buffer solution B was to wash out the impurities that were not specifically bound to the heparin affinity column. In this study, different concentrations of NaCl (0, 50, 100, 200, and 300 mmol/L) were selected for testing. e results showed that the reaction of LF was enhanced upon increasing NaCl concentrations in the range of 0-100 mmol/L (Figure 3(d)). is might be attributed to the fact that buffer B washed away nonspecifically bound e flow rate, time, and mobile phase ratio of gradient elution were appropriately adjusted according to factors such as peak output and peak shape to complete the optimization of gradient elution conditions. heterologous impurities, thereby purifying and enriching the target protein. However, the high salt concentration required to elute LF also leads to the release of the bound target protein.

Effect of NaCl Concentrations in the Elution Solution.
NaCl concentration in the elution solution was an important parameter that could influence the desorption of LF from heparin. erefore, four NaCl concentrations of 0.5 mol/L, 1.0 mol/L, 1.5 mol/L, and 2.0 mol/L were used to elute the sample. No elution peak was observed at low physiological salt concentrations, indicating potential adsorption of the protein to the column matrix (data not shown). Higher elution efficiencies of LF were observed with high salt concentrations at a constant elution volume (Figure 3(e)), and 100 mM NaCl was selected for subsequent extractions.

Method Validation
Linearity. e linearity was checked using samples of standard human LF at six different concentrations   (5-200 mg/L). e linear regression equation was determined as follows: y � 16881x + 161391 with a correlation coefficient (R 2 ) of 0.9977 (Supplementary Figure 1). (LOQ). e LOD and LOQ were 1 mg/L and 5 mg/L, which were calculated as the lowest concentrations that demonstrated signal-to-noise ratios of 3 and 10, respectively. Precision. Intermediate precision was determined, which was defined as the agreement between the test results obtained using the same method performed with identical test material in the same laboratory by the same operators using the same equipment over several days [17]. It was expressed in percentage RSD. To evaluate the intraday and interday precision of the method, a diluent of the breast milk sample was pretreated under the aforementioned optimal test conditions each day for four consecutive days. Results are summarized in Table 2.

Limit of Detection (LOD) and Limit of Quantitation
Accuracy. e accuracy of the method was evaluated using the verification method of standard addition recovery. e mean recovery rate was 103.54% (Table 3), which demonstrated that the measured values agreed well with the true values.

Milk Sample Analysis.
All collected breast milk samples were divided into two groups based on the lactation intervals and analyzed with UPLC using the current optimized method. e concentrations of HLF observed throughout the lactation period are shown in Figure 4. e results showed that lactation time significantly influenced HLF concentration, and the concentration observed during lactation time of ≤15 days was significantly higher than that observed with lactation time >15 days (p < 0.05).

Discussion
e efficiency of heparin-binding affinity was demonstrated in conjunction with the purification by measuring the recovery of standard LF loaded on the column. e recovery was determined to be nearly 100%, confirming the efficiency of this method. ese results indicate that the column was suitable for the isolation and purification of HLF.
In the present study, we demonstrated an optimized method for LF purification employing a heparin affinity column under specific conditions that preserved LF stability. e details of the purification process are as follows: LF was effectively extracted from samples with 100 mmol/L phosphate buffer at pH 5.0, retained by the heparin affinity column, and eluted using a phosphate buffer with 1.0 mol/L NaCl. e choice of the chromatographic column is also crucial for the separation of HLF from other components of the sample. C4 and C18 silica gel columns could be used to separate LF during the selection of chromatographic columns. With more than 20 injections, the C18 column often exhibits excessive baseline noise and spikes and poor reproducibility of the retention time [17]. e C4 column demonstrates good peak shape, high recovery, and low levels of protein adsorption, which is suitable for protein analysis [25]. As a result, the widely used C4 column was selected for higher stability and separation efficiency.
Notably, thermal denaturation of LF affected the binding affinity of LF and heparin [17]. Our method estimated the concentration of native LF in human milk. ermally denatured LF could have been eluted as the sample matrix.
Affinity chromatography is the predominant unit operation performed during large-scale bioprocessing and purification of therapeutic proteins. Affinity chromatography has   been widely used for the purification of native LF because of its high binding specificity with the target protein. ELISA can be used to detect the native form of proteins with sensitivity at picogram and subpicogram concentrations. e concentration of HLF observed in the present study was similar to that found by Paulaviciene et al. [29], who reported 2.5 ± 1.07 mg/ mL during 14-16 days. Although ELISA has high sensitivity and versatility, its reproducibility is relatively poor. Hence, few studies have examined the detection of LF by this method. is method did not only detect active LF attributed to the substrate specificity of heparin but also demonstrated better reproducibility than ELISA.

Conclusions
Here, we described a simple and rapid strategy in which a heparin affinity column was used for the purification of a complex sample containing LF before LF analysis. For optimal purification, several factors that may influence the binding efficiency were investigated. e method was validated in terms of accuracy, precision, LOD, LOQ, and linearity.
e results demonstrate the potential for commercial applications of this method for LF analysis.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that there are no conflicts of interest regarding the publication of this paper.

Authors' Contributions
e first and second authors contributed equally to this manuscript.