See “Bovine Lactoferrin, Human Lactoferrin, and Bioactivity” by Bhatia on page 589.
Lactoferrin (LF) is a multifunctional protein involved in many biological functions, ranging from antibacterial and antiviral activities, facilitating iron (Fe) absorption, involvement in immune function, and bone growth to enhancement of cellular proliferation and differentiation (1–3). It is found in extracellular secretions such as milk, pancreatic fluid, bile, saliva, and mucosal excretions, but it also is present in neutrophils, where it participates in phagocytotic killing. It is particularly abundant in human colostrum and milk, with concentrations ranging from 1 to 6 g/L (3). It is also present in cow's milk, albeit at considerably lower concentrations (0.03–0.1 g/L in mature bovine milk). Because of its many biological activities, purified bovine LF (bLF) is commercially available (CbLF) in large quantities for various applications.
Human LF (hLF) has been shown to bind to a specific LF receptor (LFR) present in the apical membrane of the small intestine (4). The receptor has been characterized, cloned, and sequenced (5), and found to be identical to a carbohydrate-binding protein in the intestine called intelectin (6). This receptor is expressed in particularly high levels during the fetal-neonatal period (7). Because both hLF and bLF are relatively resistant against digestion by proteolytic enzymes (8,9), and hLF has been found in intact form at substantial amounts in the stool of exclusively breast-fed infants (9), it is possible that milk LF resisting digestion can bind to the intestinal LFR.
The biological activities of LF provided by the diet (eg, breast milk, bLF-fortified infant formulas) can be divided into 2 groups: local effects of LF in the gut lumen, such as bacteriostatic/bactericidal effects; and systemic effects, mediated by the LFR, for example, Fe uptake, immunomodulatory effects, and epithelial growth and differentiation. The latter effects would be caused by either LF being internalized by the LFR or LF binding to the membrane-bound receptor causing cell-signaling events. We have shown that hLF can be internalized by human intestinal epithelial cell model (Caco-2 cells) and that both the protein and Fe are taken up (5). hLF has been shown to bind to DNA and affect gene transcription (10–12) and expression of interleukin (IL)-18 (13) and transforming growth factor-β1(TGF-β1) (Jiang and Lönnerdal, unpublished data). The effects of LF on proliferation and differentiation of Caco-2 cells appear to depend on LF concentration. LF in high concentration stimulates cell proliferation, whereas LF in low concentration enhances differentiation (14). Proliferation and differentiation of intestinal epithelial cells are critical parts of intestinal development, which are not only essential in nutrient absorption in early life but also are important to the responsiveness of the intestine to physiological or pathological challenges in later life (15).
In our earlier studies, we showed low binding of bLF to the intestinal LFR (4). A 2008 study, however, suggested that bLF actually can bind to human intestinal cell monolayers (16). It has been demonstrated that Fe-free and Fe-saturated LF have different physiological functions (17,18). We therefore optimized the LF isolation method and investigated whether native own purified bLF and CbLF with a naturally low degree of Fe saturation (native-LF [N-LF]) or in the Fe-saturated form (holo-LF [H-LF]) can bind to the LFR, become internalized, and affect cell growth and cytokine expression. We also explored the ability of these forms of bLF to resist digestion in vitro and compared the results with those obtained for hLF.
PATIENTS AND METHODS
Preparation of LFs
hLF was isolated from pooled frozen human milk from various donors. The milk was first centrifuged at 8500g for 20 minutes to remove fat. Casein was then removed by the addition of CaCl2 to 60 mmol/L, adjustment of pH to 4.6 with HCl, and then centrifugation at 8500g for 20 minutes. The whey was then run through a column with Heparin-Sepharose 6 Fast Flow resin (GE Healthcare, Piscataway, NJ) in running buffer (0.05 mol/L Tris HCl, pH 8.0). The column was washed with running buffer and LF was eluted by the same buffer containing 2 different concentrations of NaCl (0.3 or 1.0 mol/L) sequentially. Five eluted fractions for each elution buffer were concentrated by ultrafiltration using a centrifugal filter unit (50K MWCO; Millipore, Bedford, MA). To examine the purity and concentration of LF in each eluted fraction, each sample (10 μL) was subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and gels were stained with Coomassie Brilliant Blue R-250 (Sigma, St Louis, MO). Based on gel results, only LF isolated from fraction 2 with an elution buffer containing the higher concentration of NaCl (1.0 mol/L) was used for subsequent experiments. bLF was isolated from fresh cow's milk from the Dairy Teaching and Research Facility, University of California, UC Davis, using the same method. CbLF was obtained from FrieslandCampina Domo (Amersfoort, the Netherlands) and dissolved in phosphate buffered saline (PBS) at a concentration similar to that of concentrated hLF and bLF.
Preparation of Fe-Saturated LFs
The degree of Fe saturation of the isolated hLF and bLF as well as CbLF was examined spectrophotometrically at optical density 280/465 using the method of Nuijens et al (19). Briefly, A280 and A465 were measured in diluted and concentrated samples, respectively. The Fe saturation ratio was represented as a percentage of Fe-saturated LF concentration (in milligrams per milliliter, A465/0.058 to total LF concentration) (19). The Fe-saturated LF (H-LF) was prepared using a method described previously with slight changes (20). Briefly, all of the N-LF samples were incubated with freshly made ferric nitrilotriacetate solution in a 4-fold molar excess to LF and allowed binding of Fe to LF overnight at 4°C. Excess Fe was then removed by centrifugation (3000g) for 90 minutes using Centriprep YM-30 (Millipore) and 3 washes using PBS followed by centrifugation (3000g) for 90 minutes. Fe saturation was then evaluated as described above.
Detection of LPS in LFs
Lipopolysaccharide (LPS) contamination of LF samples (1 mg/mL in PBS) was examined using a Limulus Amebocyte Lysate kit (QCL-1000; Lonza, Walkersville, MD) according to the manufacturer's instruction. Three different infant formulas were purchased locally (Davis, CA), reconstituted according to manufacturers’ instructions, and LPS was measured in these formulas.
In Vitro Digestion of LFs
The gastric digestion phase was started by adjusting each LF sample (1 mg/mL in PBS) to pH 2.0 or 4.0 with HCl (1 mol/L). Porcine pepsin (2% in 1 mmol/L HCl, 1:12.5 ratio of pepsin to protein; Sigma) was added and samples were shaken in an incubator shaker (120–140 rpm; New Brunswick Scientific, Edison, NJ) for 30 minutes at 37°C. After the gastric incubation, the pH value of the samples was adjusted to 7.0 with NaHCO3 (1 mol/L) and pancreatin (0.4% in 0.1 mol/L NaHCO3, 1:62.5 ratio of pancreatin to protein; Sigma) was added. The samples were then further incubated in an incubator shaker (120–140 rpm) for 30 minutes at 37°C. After the incubation, pancreatin was inactivated in an 85°C water bath for 3 minutes. To evaluate the effects of pH changes and 37°C incubation alone on LFs, samples without enzyme treatments were analyzed at the same time. LF and the in vitro digestion samples generated from proteolysis (20 μg/lane) were boiled for 5 minutes in Laemmli sample buffer (Bio-Rad, Hercules, CA) with β-mercaptoethanol (5%) and then subjected to SDS-PAGE and gels were stained with Coomassie Brilliant Blue R-250 (Sigma). To evaluate whether proteins in formula affect digestion of LFs, powdered infant formula protein (13 g/L, Mead Johnson Nutritionals, Evansville, IN) was added to N-LFs and then in vitro digestion (pH 4.0, porcine pepsin and pancreatin sequential digestion) was performed as described above.
LF samples (20 μg/lane) were separated by SDS-PAGE and gels were transferred to nitrocellulose membranes. The membrane was incubated in blocking buffer (5% bovine serum albumin in PBST:0.1% Tween-20 in PBS) for 45 minutes at room temperature, washed 3 times with PBST, and probed with primary antibodies (goat anti-bLF, 1:1000, Bethyl Laboratories, Montgomery, TX; rabbit anti-hLF 1:500, Abcam, Cambridge, MA) in blocking buffer for 45 minutes at room temperature. After 3 washes with PBST, primary antibodies were detected with horseradish peroxidase–conjugated donkey anti-goat immunoglobulin G (1:20,000; Santa Cruz Biotechnology, Santa Cruz, CA) or horseradish peroxidase–conjugated donkey anti-rabbit immunoglobulin G (1:20,000; Amersham Pharmacia Biotech, Piscataway, NJ) in blocking buffer for 45 minutes. Bands were visualized with West Femto substrate (Pierce, Rockford, IL) after a final wash (30-minute wash with PBST).
Human colon adenocarcinoma Caco-2 cells (HTB-37; American Type Culture Collection, Rockville, MD) were maintained in a humidified incubator at 37°C under an atmosphere of 5% CO2 in Minimum Essential Medium (Life Technologies Inc, Gaithersburg, MD) supplemented with fetal bovine serum (FBS, 10%), streptomycin (100 μg/mL), and penicillin (100 U/mL; Sigma). Cells between passages 20 to 55 were harvested at 90% confluence and the medium was changed every other day. In a pilot study on LFR expression in Caco-2 cells during a 25-day culture period, LFR expression was shown to be at a maximum on day 16. Therefore, day 16 Caco-2 cells were used for our experiments.
Binding and Uptake of LF Protein and LF-bound Fe
All of the LF samples were labeled with 125I (Amersham, Buckinghamshire, UK) using the Iodogen (Pierce) method. To examine binding and uptake of LF-bound Fe, all of the LF samples were labeled with 59Fe using a method described previously (21). Caco-2 cells on 24-well plates were incubated in serum-free medium containing 125I-labeled LF samples (50,000 cpm/well) or 59Fe-labeled LF samples (50,000 cpm/well) for 120 minutes at 4°C (binding) or 37°C (uptake). Cells were rinsed with ice-cold PBS 3 times and digested with NaOH (1 mol/L). After cells were solubilized, cell-associated radioactivity was quantified in a gamma counter (Packard Minaxi-γ; Meriden, CT). Nonspecific binding was assessed by incubation with unlabeled LF (100-fold molar excess) and subtracted from total binding or uptake, resulting in specific binding and uptake, respectively.
Effects of LFs on Cellular Proliferation and Differentiation
To determine the effects of LFs on cellular proliferation of intestinal epithelial cells, Caco-2 cells were grown in 96-well plates until approximately 50% confluence. Cells were incubated in cell culture medium with FBS (2%, n = 6 samples/treatment) containing LFs (200 μg/mL) for 24 hours, and the effects of LFs on cell proliferation were evaluated using an enzyme-linked immunosorbent assay bromodeoxyuridine (BrdU) kit (Roche, Indianapolis, IN) according to the manufacturer's instructions. To investigate the effects of N- and H-LF on cellular differentiation of intestinal epithelial cells, Caco-2 cells were grown in 24-well cell culture plates until they reached 90% confluence. LFs (50 μg/mL) were added (n = 6 samples/treatment) to cells in culture medium containing FBS (2%) and incubated for 72 hours at 37°C. Cell lysates were then prepared in cell homogenization buffer (20 mmol/L Tris-HCl, pH 7.5; 150 mmol/L NaCl; 1 mmol/L ethyleneglycol tetracetic acid 1% Triton X-100; 1 × Roche complete protease inhibitor cocktail). Alkaline phosphatase activity in cell lysates was measured as an indicator of cellular differentiation (22) using p-nitrophenyl phosphate as a colorimetric phosphatase substrate. BrdU and alkaline phosphatase activities were normalized to total cell protein, which was measured by the Bradford protein assay.
Effects of LFs on IL-18 Secretion
After Caco-2 cells on 24-well plates were treated with LFs (400 μg/mL) for 72 hours at 37°C, the cell culture media were collected and concentrated with an YM-10 Ultra-10K device (Millipore). The secreted IL-18 was measured using an IL-18 enzyme-linked immunosorbent assay kit (Bender MedSystems Inc, Burlingame, CA) following the manufacturer's instructions.
Effects of LFs on TGF-β1 Expression
Caco-2 cells (day 8) on 6-well plates were treated with LFs (50 μg/mL) for 72 hours at 37°C. Total RNA was then extracted using Trizol (Invitrogen, Carlsbad, CA) according to the manufacturer's instruction. RNA integrity was evaluated by electrophoresis using agarose gel (1%) stained by ethidium bromide (Sigma). cDNA was generated with RNA (1 μg) using a high-capacity cDNA reverse transcription kit (Applied Biosystems, Foster City, CA) following the manufacturer's instructions. Gene-specific primers to TGF-β1 (5′-AAATTGAGGGCTTTCGCCTTA-3′ and 5′-GAACCCGTTGATGTCCACTTG-3′) and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (5′-CCTCCCGCTTCGCTCTCT-3′ and 5′-GGCGACGCAAAAGAAGATG-3′) were selected using Primer Express Software (Applied Biosystems) and ordered from Operon Technologies (Alameda, CA). Real-time polymerase chain reaction was performed on the cDNA reaction mixture (2 μL) and SYBR Green (Applied Biosystems) using the ABI 7900 HT real-time thermocycler (Applied Biosystems). The cycling parameters were 50°C for 2 minutes, 95°C for 10 minutes, 40 cycles including 95°C for 15 seconds, 60°C for 1 minute, and a final step at 60°C for 15 seconds and 95°C for 15 seconds, respectively. The linearity of the dissociation curve was analyzed using the ABI 7900 HT software, and the mean cycle time of the linear part of the curve was designated as Ct. Each sample was analyzed in triplicate and normalized to GAPDH using the following equation: fold-change = 2(Ct Gene − Ct GAPDH). Values are shown as mean fold change ± standard deviation, relative to control (set to 1).
Data represent mean ± standard deviation from 2 to 3 independent experiments. Comparisons between treatment and control were conducted using an unpaired Student t test or 1-way analysis of variance (Prism GraphPad, Berkeley, CA). P < 0.05 was considered to be statistically significant.
Evaluation of Fe Saturation of hLF, bLF, and CbLF
hLF had a low degree of Fe-saturation, 10.9%, as assessed spectrophotometrically. bLF isolated from cow's milk in our laboratory and CbLF were somewhat more Fe saturated at 13.2% and 17.1%, respectively (Table 1). All the 3 LF samples were capable of binding Fe, as assessed by adding freshly made ferric nitrilotriacetate solution (20) and ultrafiltering away any excess Fe. The increase in A465 reflected complete Fe saturation, that is, formation of H-LF.
Detection of LPS Contamination in hLF, bLF, and CbLF
A high-sensitivity LPS assay, Limulus Amebocyte Lysate Assay, was used to examine the LPS contamination of LF samples. hLF and bLF purified at our laboratory contained LPS albeit at low concentrations (1.0 and 2.2 endotoxin units/mL respectively; Fig. 1; 1 endotoxin unit/mL = 0.2 ng/mL), whereas CbLF contained LPS at an even lower concentration (0.80 endotoxin units/mL). The 3 commercial infant formulas contained higher levels of LPS, 3.29 to 5.01 endotoxin units/mL.
In Vitro Proteolysis
Gel electrophoresis demonstrated that CbLF contains some other proteins, with molecular weights (MW) lower than that of LF (approximately 80 kD) (Fig. 2A). These proteins appear to be fragments of LF, as shown by Western blotting, with an antibody against hLF or bLF (Fig. 2B). The purified hLF and bLF also contained small amounts of proteins with MWs lower than that of LF, which may be other contaminating proteins as indicated by the lack of staining in the Western blot.
Treatment of the bLFs and CbLFs with pH 2.0 and adjustment back to pH 7.0 (Fig. 2C I and 2E I) resulted in a slight increase in the amount of proteins with other MW than LF as compared with treatment at pH 4.0 (Fig. 2C IV and 2E IV), possibly because of dissociation of high-MW protein complexes. It should be noted that the Western blots were somewhat overexposed to show minor amounts of LF fragments.
Treatment of the LFs with pepsin at pH 2.0 followed by pancreatin digestion resulted in the complete disappearance of intact LF (Fig. 2C III, 2D III, and 2E III). When the pepsin digestion was performed at pH 4.0, a major part of hLF remained intact, whereas only a part of N-bLF and N-CbLF were intact, with some fragments with MWs of approximately 43 kD being formed (Fig. 2C V, 2D V and 2E V). After subsequent pancreatin digestion, most of the hLF was digested, whereas the bLFs showed patterns similar to those observed after pepsin digestion. The LF fragments of the bLFs formed by pepsin digestion remained intact after pancreatin digestion (Fig. 2C VI and 2E VI). H-hLF remained intact after pepsin digestion at pH 4.0 followed by pancreatin digestion, whereas the H-bLFs were completely digested (Fig. 2C VI, 2D VI and 2E VI). To evaluate whether proteins in formula affect the resistance of LFs to in vitro digestion, infant formula proteins in powdered form were added to solutions of N-LFs, and in vitro digestion was performed using pepsin and pancreatin sequentially at pH 4.0 (Fig. 3). No significant differences in LF digestibility were found when infant formula proteins were added to LFs, suggesting that LF added to infant formula has the capacity to resist digestion to some extent. A longer exposure time for the Western blots was used in this experiment, as shown in Figure 2. The results clearly show that some amounts of both intact LFs and larger fragments of LF resist digestion.
Uptake of LF Protein and LF-bound Fe by the Human Intestinal LFR
All of the LF samples bound to Caco-2 cells at 4°C. For LF protein binding, based on 125I-labeling, no differences were observed between N- and H-forms for all of the LFs (Fig. 4A), but lower binding was shown for both our own purified bLF and cbLF compared with hLF (Fig. 4A), and our own purified bLFs had the lowest binding ability. Binding of LF-bound Fe of both our own purified bLFs and cbLFs was lower than that of H-hLF and binding of LF-bound Fe of H-bLF was lowest among the H-forms of LFs, but no significant differences were found between N- and H-forms of the LF samples (Fig. 4B). Both LF protein (125I-labeled) and LF-bound Fe (59Fe labeled) were well taken up from hLF by Caco-2 cells (Fig. 5). In comparison with hLF, significantly less LF was taken up from bLFs and CbLFs, and bLFs had the lowest uptake among all of the LFs (Fig. 5A). There was significantly less Fe taken up from bLFs and CbLFs compared with hLFs, and similar results were found for H-bLFs and H-CbLFs (Fig. 5B). Interestingly, both uptake and binding of LF-bound 59Fe were significantly higher than those of 125I-labeled protein, suggesting that the mechanism by which Fe is taken up from LF is different from that by which LF protein is bound and taken up.
Effects of LFs on Cell Proliferation and Differentiation
N-hLF, N-bLF, and N-CbLF significantly enhanced cell proliferation as measured by the BrdU assay (46%, 30%, and 22%, respectively) (Fig. 6A). N-hLF, N-bLF, and N-CbLF also significantly enhanced cell differentiation as measured by increased alkaline phosphatase activity (40%, 27%, and 32%, respectively) (Fig. 6B). The H-forms of the LFs did not have any effect on cellular proliferation and differentiation.
Effects of LF on IL-18 Secretion and TGF-β1 Expression
N-forms of LF except CbLF did not have effects on IL-18 secretion and all of the H-forms of LFs dramatically enhanced IL-18 secretion. The effect of CbLF was significantly higher than that of bLF and hLF (Fig. 7A).
N-forms of hLF and bLF significantly stimulated expression of TGF-β1, but this effect was less pronounced for CbLF. H-forms of the LFs did not show any significant effects on expression of TGF-β1 (Fig. 7B).
All 3 LFs investigated had a relatively low degree of Fe saturation. hLF had the lowest Fe saturation, 10.9%, which is similar to what has been shown (23). In fact, even if all of the Fe in human milk would be bound to LF, only a fraction of its Fe-binding capacity would be used because there is a large molar excess of LF as compared with Fe. Our isolated bLF was saturated to 13.2% and cbLF to 17.1%. Thus, all 3 forms are largely present in the Fe-free form (apo-form). Because the Fe concentration of human and bovine milk is low, it is unlikely that LF Fe saturation would change in the gastrointestinal tract following consumption by an infant or a child, unless another food with Fe was ingested at the same time or if Fe drops were added. One likely application for commercial bLF, however, is to add it to infant formula, which is always fortified with Fe at relatively generous levels. Our results clearly showed that all 3 forms of LF were capable of binding Fe in a stoichiometric ratio (2 Fe/LF molecule) and being transformed to the H-form, suggesting a maintained native tertiary structure. Therefore, if bLF or a fraction thereof remains intact or in large fragments in the gastrointestinal tract, there are 2 possible scenarios with regard to its Fe saturation. If CbLF is added to powdered formula (the most common type of formula in most countries), the time for LF to bind Fe is relatively limited (dilution with water followed by consumption and passage through the stomach/upper small intestine) and apo-LF is unlikely to be transformed to H-LF because the binding of Fe to LF is a complicated event (24). If bLF is added to liquid formula (the most common type in the United States), however, generous amounts of Fe in the product are likely to bind to apo-LF with time and predictably most, if not all, LF will be in the H-form, depending on the level of LF and Fe added to the formula. It is therefore likely that the biological activities of bLF will be different depending upon the type of formula to which it is being added.
It is well known that LF binds LPS strongly (25). In fact, part of the bactericidal activity of LF is believed to be because of the capacity of LF to bind and remove LPS from the outer membrane of pathogenic bacteria (26). Thus, the ability of LF to kill bacteria by its LPS-binding capacity may be an important part of the defense against infection. CbLF frequently has been contaminated with LPS (27), presumably picked up by bLF during the isolation process. If this occurs to a considerable degree, it is possible that the bactericidal activity of LF, or other functions, potentially dependent upon unoccupied LPS-binding sites, may be compromised. We therefore investigated the LPS content of the 3 LFs included in our study. A new, sensitive Limulus Amebocyte Lysate Assay for LPS was used and we found low concentrations of LPS in all 3 forms of LF, particularly in the CbLF (0.80 endotoxin units/mL). Additionally, we evaluated LPS levels of several commercial infant formulas using the same method. LPS levels in these infant formulas were in the range of 3.29 to 5.01 endotoxin units/mL, which is low but higher than that for all of the LF samples. It is therefore unlikely that the amounts of LPS added by including LF in formula would be physiologically significant. The small amounts of LPS that we found in the LFs may be caused by low concentrations of LPS in the water used for purification and ultrafiltration processes, but it is also possible that they are caused by minute quantities of bacteria in the milk used for LF purification. Neither human nor bovine milk is sterile, and because the binding affinity of LF for LPS is high, concentrated LF may have the capacity to pick up small amounts of LPS.
In vitro digestion of the 3 LFs at a simulated gastric pH of 4.0, combined with pancreatin digestion, showed that only H-hLF completely resisted digestion, a small amount of N-bLF and N-CbLF remained intact, and some immunoreactive fragments were seen in N-hLF, N-bLF, and N-CbLF. The H-bLFs were digested completely under these conditions. It is interesting that H-hLF is more resistant to in vitro digestion, which is probably because the H-form of hLF has a more compact conformation than the apo-form (28). The H-forms of both own purified and commercial bLFs were less resistant to in vitro digestion than the N-form, probably because of structural differences between hLF and bLF. The amino acid sequence identity of hLF and bLF is 69% according to the sequence alignment analysis (http://www.ebi.ac.uk/Tools/msa/clustalw2), and the conformation of H-hLF exhibits greater closure than that of H-bLF (29). It has been shown that gastric acid secretion is low in infants (30) and that the stomach pH remains at pH 4 to 5 for a considerable time after a meal. We have shown that hLF and recombinant hLF survive digestion at pH 3.5 (21), which is because of the remarkable stability of LF against proteolytic enzymes (8). In fact, both hLF and bLF have been found to be intact in the stool of breast-fed (9) and formula-fed (31) infants, respectively. Although our LFs were digested to a considerable level under these in vitro conditions, it should be recognized that both intact LFs and larger fragments escaped digestion when pepsin digestion was carried out at pH 4.0 both in isolated form and in the presence of infant formula proteins, as shown in Figure 3. It is possible that the fraction of LF surviving digestion is sufficient to exert bioactivities. For example, TGF-β1 expression was significantly increased at an LF concentration of 50 μg/mL, which is approximately 3% to 5% of the concentration of LF in human milk. Similarly, LF in human milk has the capacity to bind 8 times the total amount of Fe in breast milk, indicating that only a fraction of LF is needed for delivery of Fe to intestinal cells. We also exposed the LFs to a simulated stomach pH of 2.0, a pH that is commonly assumed to approximate that of the adult stomach, and pepsin and pancreatin digestion. When this lower pH was used, LF was effectively digested and only minor fragments remained; however, recent studies on human adults show that the gastric pH after a meal rarely drops below 3.5 (32,33), suggesting that a gastric pH of 2.0 may only reflect fasting conditions and that LF as part of a meal also may survive digestion in adults. Several other human milk proteins have been shown to resist proteolytic degradation in vitro, for example, α1-antitrypsin (34) and haptocorrin (35), and in vivo, for example, secretory IgA and (36), demonstrating the limited capacity for protein digestion in the infant gut.
We have shown that hLF is taken up by human intestinal Caco-2 cells in culture (37) and that this is facilitated by a clathrin-mediated endocytotic process (38). In the present study, we showed that both our own purified bLF and cbLF were internalized by Caco-2 cells and that Fe was internalized as well. As discussed previously, we are uncertain why bLF did not bind to the LFR in our earlier studies, but the purity/quality of the bLF is likely a reason. Our binding and uptake results are in agreement with the recent binding study of Shin et al (16), who showed that the recombinant hLFR (intelectin) binds to microtiter plates coated with hLF or bLF. Interestingly, according to the study of Shin et al, LF peptides from a pepsin hydrolysate of bLF also bound to the recombinant hLFR but to a lower degree. It is not yet known whether these peptides also can be internalized by epithelial cells. The uptake of bLF and CbLF and its bound Fe strongly suggests that the human intestinal LFR recognizes bLF in the N-form. Uptake of bLF was lower than that of hLF, which is not surprising because their tertiary structures differ, which may affect the internalization process. It is difficult to evaluate whether this short-term quantitative difference also would affect the uptake of bLF in infants consuming formula with bLF and when the intestinal mucosa is exposed to bLF in every meal during extended time periods. Fe binding and uptake from H-LFs were significantly higher than those from N-LFs (Figs. 3 and 4), which is similar to results obtained from human monocytes (39). All of the small peptides can be labeled with 125I because labeling reactions occur especially with the aromatic ring of amino acids (40), whereas not all of the peptides can be labeled with 59Fe, which may result in lower binding of 125I-labeled LFs than 59Fe-labeled LFs. When transferrin binds to the extracellular transferrin receptor, transferrin-bound Fe is released (41). Fe may be released after LFs bind to the LFR because of changes in LF conformation; therefore, Fe may bind to the plasma membrane or be internalized by cells through a mechanism that is different from that for LF protein.
All 3 forms of N-LF significantly increased cell proliferation. We found no effect for the H-LFs. This is similar to findings of Buccigrossi et al (14) but not to observations by Oguchi et al (17), who found that only H-LF increased cell proliferation. The latter study used differentiated Caco-2 cells as a model to examine the effects of LF on intestinal proliferation. Proliferative mechanisms are likely diminished in differentiated Caco-2 cells. It is possible that the effects of LF on differentiated cells would be different from effects on greatly proliferative, nondifferentiated cells. All 3 forms of N-LF also significantly enhanced cell differentiation, but H-LF had no effect, which is also similar to the study by Buccigrossi et al (14). The mechanism(s) behind this growth stimulatory effect is not yet known, but hLF has been shown to bind to DNA (12), which in turn may affect gene transcription. This would be consistent with observations that LF can affect the expression of IL-18 and thus affect immune function systemically, such as enhancing natural killer cell activity (42). We found a positive effect of all of the 3 H-LFs and of N-CbLF but not of N-hLF and N-bLF on IL-18 expression by Caco-2 cells. It has been shown in a mouse model that orally administered bLF induces caspase-1 and IL-18, which may affect immune function by potentiating the activity of T and natural killer cells, and also inhibition of carcinogenesis (13). It also has been shown that oral bLF increases IL-18, interferon-α, and interferon-β in Peyer patches and mesenteric lymph nodes and in turn increases natural killer cell activity (42). Moreover, we showed that N-hLF and N-bLF significantly stimulated expression the TGF-β1 gene. TGF-β1 is a pleiotropic cytokine and plays important roles in the maintenance of immune homeostasis (43). Thus, it is evident that bLF can affect immune function at multiple levels. It is unclear, however, why no enhancing IL-18 effect was observed for the N-forms of hLF and bLF in our study and no stimulatory effects on TGF-β1 for H-forms of LFs.
In conclusion, cbLF and hLF and bLF purified from fresh cow's milk in the laboratory was predominantly in the apo-form, confirmed with a still high Fe-binding capacity, whereas hLF and purified bLF contained low amounts of LPS. LPS was practically absent in cbLF. Similar to hLF, purified bLF and cbLF were bound to human intestinal epithelial cells and were internalized because Fe was bound to them. At a pH similar to that of the human infant stomach, these bLFs partly resisted proteolytic degradation or were digested to immunoreactive fragments in vitro, whereas hLF was digested completely. Thus, it is possible that cbLF can be used as a supplement for application in products for infants and that it can achieve some of the functions of hLF.
The authors appreciate constructive discussions with Dr Marjan Gros.
1. Brock JH. The physiology of lactoferrin. Biochem Cell Biol
2. Ward PP, Paz E, Conneely OM. Multifunctional roles of lactoferrin: a critical overview. Cell Mol Life Sci
3. Lonnerdal B, Iyer S. Lactoferrin: molecular structure and biological function. Annu Rev Nutr
4. Kawakami H, Lonnerdal B. Isolation and function of a receptor for human lactoferrin in human fetal intestinal brush-border membranes. Am J Physiol
5. Suzuki YA, Shin K, Lonnerdal B. Molecular cloning and functional expression of a human intestinal lactoferrin receptor. Biochemistry
6. Tsuji S, Uehori J, Matsumoto M, et al. Human intelectin is a novel soluble lectin that recognizes galactofuranose in carbohydrate chains of bacterial cell wall. J Biol Chem
7. Lopez V, Suzuki YA, Lonnerdal B. Ontogenic changes in lactoferrin receptor and DMT1 in mouse small intestine: implications for iron absorption during early life. Biochem Cell Biol
8. Brines RD, Brock JH. The effect of trypsin and chymotrypsin on the in vitro antimicrobial and iron-binding properties of lactoferrin in human milk and bovine colostrum. Unusual resistance of human apolactoferrin to proteolytic digestion. Biochim Biophys Acta
9. Davidson LA, Lonnerdal B. Persistence of human milk proteins in the breast-fed infant. Acta Paediatr Scand
10. Penco S, Scarfi S, Giovine M, et al. Identification of an import signal for, and the nuclear localization of, human lactoferrin. Biotechnol Appl Biochem
11. Penco S, Pastorino S, Bianchi-Scarra G, et al. Lactoferrin down-modulates the activity of the granulocyte macrophage colony-stimulating factor promoter in interleukin-1 beta-stimulated cells. J Biol Chem
12. He J, Furmanski P. Sequence specificity and transcriptional activation in the binding of lactoferrin to DNA. Nature
13. Iigo M, Shimamura M, Matsuda E, et al. Orally administered bovine lactoferrin induces caspase-1 and interleukin-18 in the mouse intestinal mucosa: a possible explanation for inhibition of carcinogenesis and metastasis. Cytokine
14. Buccigrossi V, de Marco G, Bruzzese E, et al. Lactoferrin induces concentration-dependent functional modulation of intestinal proliferation and differentiation. Pediatr Res
15. Drozdowski LA, Clandinin T, Thomson AB. Ontogeny, growth and development of the small intestine: understanding pediatric gastroenterology. World J Gastroenterol
16. Shin K, Wakabayashi H, Yamauchi K, et al. Recombinant human intelectin binds bovine lactoferrin and its peptides. Biol Pharm Bull
17. Oguchi S, Walker WA, Sanderson IR. Iron saturation alters the effect of lactoferrin on the proliferation and differentiation of human enterocytes (Caco-2 cells). Biol Neonate
18. Norrby K. Human apo-lactoferrin enhances angiogenesis mediated by vascular endothelial growth factor A in vivo. J Vasc Res
19. Nuijens JH, van Berkel PH, Geerts ME, et al. Characterization of recombinant human lactoferrin secreted in milk of transgenic mice. J Biol Chem
20. van Berkel PH, Geerts ME, van Veen HA, et al. Glycosylated and unglycosylated human lactoferrins both bind iron and show identical affinities towards human lysozyme and bacterial lipopolysaccharide, but differ in their susceptibilities towards tryptic proteolysis. Biochem J
21. Suzuki YA, Kelleher SL, Yalda D, et al. Expression, characterization, and biologic activity of recombinant human lactoferrin in rice. J Pediatr Gastroenterol Nutr
22. Alpers DH, Mahmood A, Engle M, et al. The secretion of intestinal alkaline phosphatase (IAP) from the enterocyte. J Gastroenterol
1994; 29 (suppl 7):63–67.
23. Hirai Y, Kawakata N, Satoh K, et al. Concentrations of lactoferrin and iron in human milk at different stages of lactation. J Nutr Sci Vitaminol (Tokyo)
24. Baker HM, Baker EN. Lactoferrin and iron: structural and dynamic aspects of binding and release. Biometals
25. Elass-Rochard E, Roseanu A, Legrand D, et al. Lactoferrin-lipopolysaccharide interaction: involvement of the 28-34 loop region of human lactoferrin in the high-affinity binding to Escherichia coli
055B5 lipopolysaccharide. Biochem J
26. Ellison RT 3rd, Giehl TJ, LaForce FM. Damage of the outer membrane of enteric Gram-negative bacteria by lactoferrin and transferrin. Infect Immun
27. Perraudin JP. Paper presented at Xth International Conference on Lactoferrin, Beijing, China, 2009
28. Grossmann JG, Neu M, Pantos E, et al. X-ray solution scattering reveals conformational changes upon iron uptake in lactoferrin, serum and ovo-transferrins. J Mol Biol
29. Moore SA, Anderson BF, Groom CR, et al. Three-dimensional structure of diferric bovine lactoferrin at 2.8 A resolution. J Mol Biol
30. Agunod M, Yamaguchi N, Lopez R, et al. Correlative study of hydrochloric acid, pepsin, and intrinsic factor secretion in newborns and infants. Am J Dig Dis
31. Spik G, Brunet B, Mazurier-Dehaine C, et al. Characterization and properties of the human and bovine lactotransferrins extracted from the faeces of newborn infants. Acta Paediatr Scand
32. Kalantzi L, Goumas K, Kalioras V, et al. Characterization of the human upper gastrointestinal contents under conditions simulating bioavailability/bioequivalence studies. Pharm Res
33. Simonian HP, Vo L, Doma S, et al. Regional postprandial differences in pH within the stomach and gastroesophageal junction. Dig Dis Sci
34. Chowanadisai W, Lonnerdal B. Alpha(1)-antitrypsin and antichymotrypsin in human milk: origin, concentrations, and stability. Am J Clin Nutr
35. Adkins Y, Lonnerdal B. Potential host-defense role of a human milk vitamin B-12-binding protein, haptocorrin, in the gastrointestinal tract of breastfed infants, as assessed with porcine haptocorrin in vitro. Am J Clin Nutr
36. Davidson LA, Lonnerdal B. Fecal alpha 1-antitrypsin in breast-fed infants is derived from human milk and is not indicative of enteric protein loss. Acta Paediatr Scand
37. Ashida K, Sasaki H, Suzuki YA, et al. Cellular internalization of lactoferrin in intestinal epithelial cells. Biometals
38. Jiang R, Lopez V, Kelleher S, et al. Apo- and holo-lactoferrin are both internalized by lactoferrin receptor via clathrin-mediated endocytosis but differentially affect ERK-signaling and cell proliferation in Caco-2 cells. J Cell Physiol
39. Birgens HS, Kristensen LO, Borregaard N, et al. Lactoferrin-mediated transfer of iron to intracellular ferritin in human monocytes. Eur J Haematol
40. Unak T, Akgun Z, Yildirim Y, et al. Self-radioiodination of iodogen. Appl Radiat Isot
41. Bali PK, Aisen P. Receptor-modulated iron release from transferrin: differential effects on N- and C-terminal sites. Biochemistry
42. Kuhara T, Yamauchi K, Tamura Y, et al. Oral administration of lactoferrin increases NK cell activity in mice via increased production of IL-18 and type I IFN in the small intestine. J Interferon Cytokine Res
43. Shi Y, Massague J. Mechanisms of TGF-beta signaling from cell membrane to the nucleus. Cell